6b244191-77a3-41ea-a314-82c6a9184b31.txt

import os
import sys
with open(sys.argv[0]) as f:
    code = f.read() # read the code of this file ASAP, for logging
import uuid
import glob
import time
import contextlib
from dataclasses import dataclass

import numpy as np
import torch
from torch import nn
import torch.nn.functional as F
import torch.distributed as dist
import torch._inductor.config as config
from torch.nn.parallel import DistributedDataParallel as DDP
# Use of FlexAttention contributed by @KoszarskyB
from torch.nn.attention.flex_attention import flex_attention, create_block_mask
flex_attention = torch.compile(flex_attention, dynamic=False)
create_block_mask = torch.compile(create_block_mask, dynamic=False)

# -----------------------------------------------------------------------------
# Muon optimizer

def zeropower_via_svd(G, steps=None):
    U, S, V = G.svd()
    return U @ V.T

@torch.compile
def zeropower_via_newtonschulz5(G, steps=10, eps=1e-7):
    """
    Newton-Schulz iteration to compute the zeroth power / orthogonalization of G. We opt to use a
    quintic iteration whose coefficients are selected to maximize the slope at zero. For the purpose
    of minimizing steps, it turns out to be empirically effective to keep increasing the slope at
    zero even beyond the point where the iteration no longer converges all the way to one everywhere
    on the interval. This iteration therefore does not produce UV^T but rather something like US'V^T
    where S' is diagonal with S_{ii}' ~ Uniform(0.5, 1.5), which turns out not to hurt model
    performance at all relative to UV^T, where USV^T = G is the SVD.
    """
    assert len(G.shape) == 2
    a, b, c = (3.4445, -4.7750,  2.0315)
    X = G.bfloat16()
    X /= (X.norm() + eps) # ensure top singular value <= 1
    if G.size(0) > G.size(1):
        X = X.T
    for _ in range(steps):
        A = X @ X.T
        B = b * A + c * A @ A # adapted from suggestion by @jxbz, @leloykun, and @YouJiacheng
        X = a * X + B @ X
    if G.size(0) > G.size(1):
        X = X.T
    return X

zeropower_backends = dict(svd=zeropower_via_svd, newtonschulz5=zeropower_via_newtonschulz5)

class Muon(torch.optim.Optimizer):
    """
    Muon - MomentUm Orthogonalized by Newton-schulz

    Muon internally runs standard SGD-momentum, and then performs an orthogonalization post-
    processing step, in which each 2D parameter's update is replaced with the nearest orthogonal
    matrix. To efficiently orthogonalize each update, we use a Newton-Schulz iteration, which has
    the advantage that it can be stably run in bfloat16 on the GPU.

    Some warnings:
    - This optimizer assumes that all parameters passed in are 2D.
    - It should not be used for the embedding layer, the final fully connected layer, or any {0,1}-D
    parameters; those should all be optimized by a standard method (e.g., AdamW).
    - To use it with 4D convolutional filters, it works well to just flatten their last 3 dimensions.
    - We believe it is unlikely to work well for training with small batch size.
    - We believe it may not work well for finetuning pretrained models, but we haven't tested this.
    - We have not yet tried this optimizer for training scenarios larger than NanoGPT (124M).

    Arguments:
        lr: The learning rate used by the internal SGD.
        momentum: The momentum used by the internal SGD.
        nesterov: Whether to use Nesterov-style momentum in the internal SGD. (recommended)
        backend: The chosen backend for the orthogonalization step. (recommended: 'newtonschulz5')
        backend_steps: The number of iteration steps to use in the backend, if it is iterative.
    """
    def __init__(self, params, lr=0.02, momentum=0.95, nesterov=True,
                 backend='newtonschulz5', backend_steps=5):
        defaults = dict(lr=lr, momentum=momentum, nesterov=nesterov, backend=backend, backend_steps=backend_steps)
        super().__init__(params, defaults)

    def step(self):

        for group in self.param_groups:

            lr = group['lr']
            momentum = group['momentum']
            zeropower_backend = zeropower_backends[group['backend']]

            # generate weight updates in distributed fashion
            total_params = sum(p.numel() for p in group['params'])
            updates_flat = torch.zeros(total_params, device='cuda', dtype=torch.bfloat16)
            curr_idx = 0
            for i, p in enumerate(group['params']):
                # luckily this will perfectly distribute a transformer with multiple of 4 layers to 8 GPUs
                if i % int(os.environ['WORLD_SIZE']) == int(os.environ['RANK']):
                    g = p.grad
                    assert g is not None
                    state = self.state[p]
                    if 'momentum_buffer' not in state:
                        state['momentum_buffer'] = torch.zeros_like(g)
                    buf = state['momentum_buffer']
                    buf.mul_(momentum).add_(g)
                    g = g.add(buf, alpha=momentum) if group['nesterov'] else buf
                    g = zeropower_backend(g, steps=group['backend_steps'])
                    g *= max(1, g.size(0)/g.size(1))**0.5
                    updates_flat[curr_idx:curr_idx+p.numel()] = g.flatten()
                curr_idx += p.numel()

            # sync updates across devices. we are not memory-constrained so can do this simple deserialization
            dist.all_reduce(updates_flat, op=dist.ReduceOp.SUM)

            # deserialize and apply updates
            curr_idx = 0
            for p in group['params']:
                g = updates_flat[curr_idx:curr_idx+p.numel()].view_as(p.data).type_as(p.data)
                p.data.add_(g, alpha=-lr)
                curr_idx += p.numel()

# -----------------------------------------------------------------------------
# PyTorch nn.Module definitions for the GPT-2 model

def norm(x):
    return F.rms_norm(x, (x.size(-1),))

class CastedLinear(nn.Linear):

    def __init__(self, in_features, out_features):
        super().__init__(in_features, out_features, bias=False)

    def forward(self, x):
        return F.linear(x, self.weight.to(x.dtype))

class Rotary(torch.nn.Module):

    def __init__(self, dim, base=10000):
        super().__init__()
        self.register_buffer('inv_freq', (1 / base) ** (torch.arange(0, dim, 2) / dim))
        self.seq_len_cached = None
        self.cos_cached = None
        self.sin_cached = None

    def forward(self, x):
        seq_len = x.shape[1]
        if seq_len != self.seq_len_cached:
            t = torch.arange(seq_len, device=x.device)
            freqs = torch.outer(t, self.inv_freq)
            self.seq_len_cached = seq_len
            self.cos_cached = freqs.cos()
            self.sin_cached = freqs.sin()
        cos, sin = self.cos_cached[None, :, None, :], self.sin_cached[None, :, None, :]
        # apply_rotary_emb(x, cos, sin)
        x1, x2 = x.chunk(2, dim=3)
        y1 = x1 * cos + x2 * sin
        y2 = x1 * (-sin) + x2 * cos
        return torch.cat((y1, y2), 3).type_as(x)

class CausalSelfAttention(nn.Module):

    def __init__(self, dim, n_head):
        super().__init__()
        assert dim % n_head == 0
        self.n_head = n_head
        self.c_q = CastedLinear(dim, dim)
        self.c_k = CastedLinear(dim, dim)
        self.c_v = CastedLinear(dim, dim)
        # value residual lambda
        self.lamb = nn.Parameter(torch.tensor(0.5)) # @Grad62304977
        # rotary embeddings
        self.rotary = Rotary(dim // n_head) # dim // n_head = head_dim
        # output projection
        self.c_proj = CastedLinear(dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x, vi, block_mask):
        B, T = x.size(0), x.size(1) # batch size, sequence length
        assert B == 1, "Must use batch size = 1 for FlexAttention"
        q = self.c_q(x).view(B, T, self.n_head, -1)
        k = self.c_k(x).view(B, T, self.n_head, -1)
        v = self.c_v(x).view(B, T, self.n_head, -1)
        v = (1 - self.lamb) * v + self.lamb * vi.view_as(v) # @Grad62304977
        q, k = norm(q), norm(k) # QK norm suggested by @Grad62304977
        q, k = self.rotary(q), self.rotary(k)
        y = flex_attention(q.transpose(1, 2), k.transpose(1, 2), v.transpose(1, 2), block_mask=block_mask)
        y = y.transpose(1, 2).contiguous().view_as(x) # re-assemble all head outputs side by side
        y = self.c_proj(y)
        return y

class MLP(nn.Module):

    def __init__(self, dim):
        super().__init__()
        self.c_fc   = CastedLinear(dim, 4 * dim)
        self.c_proj = CastedLinear(4 * dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x):
        x = self.c_fc(x)
        x = F.relu(x).square() # https://arxiv.org/abs/2109.08668v2; ~1-2% better than GELU; suggested by @SKYLINEZ007 and @Grad62304977
        x = self.c_proj(x)
        return x

class Block(nn.Module):

    def __init__(self, config):
        super().__init__()
        self.attn = CausalSelfAttention(config.n_embd, config.n_head)
        self.mlp = MLP(config.n_embd)
        self.lambdas = nn.Parameter(torch.tensor([1., 0.]))

    def forward(self, x, vi, x0, block_mask):
        x = self.lambdas[0] * x + self.lambdas[1] * x0
        x = x + self.attn(norm(x), vi, block_mask)
        x = x + self.mlp(norm(x))
        return x

# -----------------------------------------------------------------------------
# The main GPT-2 model

@dataclass
class GPTConfig:
    vocab_size : int = 50304
    n_layer : int = 12
    n_head : int = 6 # head dim 128 suggested by @Grad62304977
    n_embd : int = 768

class GPT(nn.Module):

    def __init__(self, config):
        super().__init__()

        # U-net design by @brendanh0gan
        self.num_encoder_layers = config.n_layer // 2 # Half of the layers for encoder
        self.num_decoder_layers = config.n_layer - self.num_encoder_layers # Remaining for decoder
        # Add learnable skip connection weights for decoder layers
        self.skip_weights = nn.Parameter(torch.ones(self.num_decoder_layers))

        self.transformer = nn.ModuleDict(dict(
            wte = nn.Embedding(config.vocab_size, config.n_embd),
            # token value embeddings by @KoszarskyB - inspired by @Grad62304977's value residual learning
            vte = nn.Embedding(config.vocab_size, config.n_embd*12),
            h = nn.ModuleList([Block(config) for _ in range(config.n_layer)]),
        ))
        self.lm_head = CastedLinear(config.n_embd, config.vocab_size)
        self.lm_head.weight.data.zero_() # @Grad62304977

    def forward(self, idx, target, attn_blocksize):

        docs = (idx == 50256).cumsum(0)
        def document_causal_mask(b, h, q_idx, kv_idx):
          causal_mask = q_idx >= kv_idx
          document_mask = docs[q_idx] == docs[kv_idx]
          window_mask = q_idx - kv_idx < attn_blocksize
          return causal_mask & document_mask & window_mask

        S = len(idx)
        block_mask = create_block_mask(document_causal_mask, None, None, S, S, device="cuda", _compile=True)

        # forward the GPT model itself
        x = self.transformer.wte(idx[None]) # token embeddings of shape (b, t, n_embd)
        x = norm(x) # @Grad62304977
        x0 = x
        vi = self.transformer.vte(idx[None]).chunk(12, dim=-1)

        # Store outputs for U-Net skip connections
        skip_connections = []
        # Encoder pass - process only the first half of the blocks
        for i in range(self.num_encoder_layers):
            x = self.transformer.h[i](x, vi[i], x0, block_mask)
            skip_connections.append(x)
        # Decoder pass - process the remaining blocks with weighted skip connections
        for i in range(self.num_decoder_layers):
            x = x + self.skip_weights[i] * skip_connections.pop()
            x = self.transformer.h[self.num_encoder_layers + i](x, vi[self.num_encoder_layers+i], x0, block_mask)

        x = norm(x)
        logits = self.lm_head(x)
        logits = 30 * torch.tanh(logits / 30) # @Grad62304977
        logits = logits.float()
        loss = F.cross_entropy(logits.view(-1, logits.size(-1)), target.view(-1))
        return loss

# -----------------------------------------------------------------------------
# Our own simple Distributed Data Loader

def _peek_data_shard(filename):
    # only reads the header, returns header data
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
    if header[0] != 20240520:
        print("ERROR: magic number mismatch in the data .bin file!")
        print("---> HINT: Are you passing in a correct file with --input_bin?")
        print("---> HINT: Dataset encoding changed recently, re-run data prepro or refer again to README")
        print("---> HINT: For example re-run: `python dev/data/tinyshakespeare.py`, then re-try")
        exit(1)
    assert header[1] == 1, "unsupported version"
    ntok = header[2] # number of tokens (claimed)
    return ntok # for now just return the number of tokens

def _load_data_shard(filename):
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
        assert header[0] == 20240520, "magic number mismatch in the data .bin file"
        assert header[1] == 1, "unsupported version"
        ntok = header[2] # number of tokens (claimed)
        # the rest of it are tokens, stored as uint16
        tokens = np.frombuffer(f.read(), dtype=np.uint16)
    assert len(tokens) == ntok, "number of tokens read does not match header?"
    return tokens

class DistributedDataLoader:
    def __init__(self, filename_pattern, T, process_rank, num_processes):
        self.process_rank = process_rank
        self.num_processes = num_processes
        self.T = T

        # glob files that match the pattern
        self.files = sorted(glob.glob(filename_pattern))
        assert len(self.files) > 0, f"did not find any files that match the pattern {filename_pattern}"

        # load and validate all data shards, count number of tokens in total
        ntok_total = 0
        for fname in self.files:
            shard_ntok = _peek_data_shard(fname)
            assert shard_ntok >= num_processes * T + 1
            ntok_total += int(shard_ntok)
        self.ntok_total = ntok_total

        self.reset()

    def reset(self):
        self.current_shard = -1
        self.advance()

    def advance(self): # advance to next data shard
        self.current_shard = (self.current_shard + 1) % len(self.files)
        self.current_position = self.process_rank * self.T
        self.tokens = _load_data_shard(self.files[self.current_shard])

    def next_batch(self):
        batch_size = self.T * self.num_processes
        buf = self.tokens[self.current_position:self.current_position+self.T+1]
        buf = torch.tensor(buf.astype(np.int32), dtype=torch.long)
        x = buf[:-1] # inputs
        y = buf[1:] # targets
        # advance current position and load next shard if necessary
        self.current_position += batch_size
        if self.current_position + batch_size >= len(self.tokens):
            self.advance()
        return x.cuda(), y.cuda()

# -----------------------------------------------------------------------------
# int main

@dataclass
class Hyperparameters:
    # data hyperparams
    input_bin : str = 'data/fineweb10B/fineweb_train_*.bin' # input .bin to train on
    input_val_bin : str = 'data/fineweb10B/fineweb_val_*.bin' # input .bin to eval validation loss on
    # optimization hyperparams
    batch_size : int = 8 # batch size, in sequences, across all devices
    sequence_length : int = 64*1024 # sequence length, in tokens
    num_iterations : int = 1530 # number of iterations to run
    warmup_iters : int = 0
    cooldown_iters : int = 600 # number of iterations of linear warmup/cooldown for triangular or trapezoidal schedule
    weight_decay : float = 0
    # evaluation and logging hyperparams
    val_loss_every : int = 125 # every how many steps to evaluate val loss? 0 for only at the end
    val_tokens : int = 10485760 # how many tokens of validation data? it's important to keep this fixed for consistent comparisons
    save_every : int = 0 # every how many steps to save the checkpoint? 0 for only at the end
args = Hyperparameters()

# set up DDP (distributed data parallel). torchrun sets this env variable
assert torch.cuda.is_available()
dist.init_process_group(backend='nccl')
ddp_rank = int(os.environ['RANK'])
ddp_local_rank = int(os.environ['LOCAL_RANK'])
ddp_world_size = int(os.environ['WORLD_SIZE'])
device = f'cuda:{ddp_local_rank}'
torch.cuda.set_device(device)
print(f"using device: {device}")
master_process = (ddp_rank == 0) # this process will do logging, checkpointing etc.

# begin logging
logfile = None
if master_process:
    run_id = str(uuid.uuid4())
    logdir = 'logs/%s/' % run_id
    os.makedirs(logdir, exist_ok=True)
    logfile = 'logs/%s.txt' % run_id
    # create the log file
    with open(logfile, "w") as f:
        # begin the log by printing this file (the Python code)
        f.write(code)
        f.write('='*100 + '\n')
def print0(s, logonly=False):
    if master_process:
        with open(logfile, "a") as f:
            if not logonly:
                print(s)
            f.write(s+'\n')
# log information about the hardware/software environment this is running on
# and print the full `nvidia-smi` to file
print0(f"Running pytorch {torch.version.__version__} compiled for CUDA {torch.version.cuda}\nnvidia-smi:")
import subprocess
result = subprocess.run(['nvidia-smi'], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
print0(f'{result.stdout}', logonly=True)
print0('='*100, logonly=True)

# convenience variables
T = args.sequence_length
# calculate the number of steps to take in the val loop.
assert args.val_tokens % (T * ddp_world_size) == 0
val_steps = args.val_tokens // (T * ddp_world_size)
# calculate the steps of gradient accumulation required to attain the desired global batch size.
assert args.batch_size % (ddp_world_size) == 0
train_accumulation_steps = args.batch_size // ddp_world_size

# load tokens
train_loader = DistributedDataLoader(args.input_bin, T, ddp_rank, ddp_world_size)
val_loader = DistributedDataLoader(args.input_val_bin, T, ddp_rank, ddp_world_size)
print0(f"Training DataLoader: total number of tokens: {train_loader.ntok_total} across {len(train_loader.files)} files")
print0(f"Validation DataLoader: total number of tokens: {val_loader.ntok_total} across {len(val_loader.files)} files")
print0('='*100, logonly=True)
x, y = train_loader.next_batch()

# there are only 50257 unique GPT-2 tokens; we extend to nearest multiple of 128 for efficiency. suggested to me by @Grad62304977.
# this originates from Karpathy's experiments.
num_vocab = 50304
model = GPT(GPTConfig(vocab_size=num_vocab, n_layer=12, n_head=6, n_embd=768))
model = model.cuda().bfloat16()
for m in model.modules():
    if isinstance(m, CastedLinear):
        m.float()
if hasattr(config, "coordinate_descent_tuning"):
    config.coordinate_descent_tuning = True # suggested by @Chillee
model = torch.compile(model)
# here we wrap model into DDP container
model = DDP(model, device_ids=[ddp_local_rank])
raw_model = model.module # always contains the "raw" unwrapped model

# init the optimizer(s)
optimizer1 = torch.optim.Adam([raw_model.transformer.wte.weight, raw_model.transformer.vte.weight], lr=0.6, betas=(0.8, 0.95), fused=True)
optimizer2 = torch.optim.Adam([raw_model.lm_head.weight], lr=0.008, betas=(0.8, 0.95), fused=True)
params = list(raw_model.transformer.h.parameters())
matrix_params = [p for p in params if p.ndim == 2]
scalar_params = [p for p in params if p.ndim < 2] + [raw_model.skip_weights]
optimizer3 = Muon(matrix_params, lr=0.05, momentum=0.95)
optimizer4 = torch.optim.Adam(scalar_params, lr=0.04, betas=(0.8, 0.95), fused=True) # note that this learning rate is neither sensitive nor tuned
optimizers = [optimizer1, optimizer2, optimizer3, optimizer4]
# learning rate decay scheduler (linear warmup and cooldown)
def get_lr(it):
    assert it <= args.num_iterations
    # 1) linear warmup for warmup_iters steps
    if it < args.warmup_iters:
        return (it+1) / args.warmup_iters
    # 2) constant lr for a while
    elif it < args.num_iterations - args.cooldown_iters:
        return 1.0
    # 3) linear cooldown
    else:
        decay_ratio = (args.num_iterations - it) / args.cooldown_iters
        return decay_ratio
schedulers = [torch.optim.lr_scheduler.LambdaLR(opt, get_lr) for opt in optimizers]

# Start training loop
training_time_ms = 0
# start the clock
torch.cuda.synchronize()
t0 = time.time()
# begin training
for step in range(args.num_iterations + 1):
    last_step = (step == args.num_iterations)
    # This effectively ignores timing first 10 steps, which are slower for weird reasons.
    # Alternately, and slightly more correctly in terms of benchmarking, we could do 10
    # steps with dummy data first, and then re-initialize the model and reset the loader.
    if step == 10:
        training_time_ms = 0
        t0 = time.time()
    timed_steps = float('nan') if step <= 11 else (step - 10) + 1 # <= 11 to avoid bug in val

    # Set the attention blocksize for the current step, in chunks of 64. By @fernbear.bsky.social
    attn_blocksize = torch.tensor(64*((step/args.num_iterations * (1792 - 64) + 64)//64), dtype=torch.int, device='cuda')

    # once in a while evaluate the validation dataset
    if (last_step or (args.val_loss_every > 0 and step % args.val_loss_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # run validation batches
        model.eval()
        val_loader.reset()
        val_loss = 0.0
        for _ in range(val_steps):
            with torch.no_grad():
                x_val, y_val = val_loader.next_batch()
                val_loss += model(x_val, y_val, attn_blocksize=attn_blocksize)
        dist.all_reduce(val_loss, op=dist.ReduceOp.AVG)
        val_loss /= val_steps
        # log val loss to console and to logfile
        print0(f'step:{step}/{args.num_iterations} val_loss:{val_loss:.4f} train_time:{training_time_ms:.0f}ms step_avg:{training_time_ms/(timed_steps-1):.2f}ms')
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    if master_process and (last_step or (args.save_every > 0 and step % args.save_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # save the state of the training process
        log = dict(step=step, code=code, model=raw_model.state_dict(), optimizers=[opt.state_dict() for opt in optimizers])
        torch.save(log, 'logs/%s/state_step%06d.pt' % (run_id, step))
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    # bit confusing: we want to make sure to eval on 0th iteration
    # but also after the very last iteration. so we loop for step <= num_iterations
    # instead of just < num_iterations (one extra due to <=), only to do
    # the validation/sampling one last time, and then we break right here as we're done.
    if last_step:
        break

    # --------------- TRAINING SECTION BEGIN -----------------
    model.train()
    for i in range(1, train_accumulation_steps+1):
        ctx = model.no_sync() if i < train_accumulation_steps else contextlib.nullcontext()
        with ctx: # there's no need to sync gradients every accumulation step
            # forward pass
            loss = model(x, y, attn_blocksize=attn_blocksize)
            # advance the dataset for the next batch
            x, y = train_loader.next_batch()
            # backward pass
            loss.backward()
        train_loss = loss.detach()
    for p in model.parameters():
        p.grad /= train_accumulation_steps
    # momentum warmup for Muon
    frac = min(step/300, 1)
    optimizer3.param_groups[0]['momentum'] = (1 - frac) * 0.85 + frac * 0.95
    # step the optimizers and schedulers
    for opt, sched in zip(optimizers, schedulers):
        opt.step()
        sched.step()
    # null the gradients
    model.zero_grad(set_to_none=True)
    # --------------- TRAINING SECTION END -------------------
    # everything that follows now is just diagnostics, prints, logging, etc.

    #dist.all_reduce(train_loss, op=dist.ReduceOp.AVG) # all-reducing the training loss would be more correct in terms of logging, but slower
    approx_time = training_time_ms + 1000 * (time.time() - t0)
    print0(f"step:{step+1}/{args.num_iterations} train_loss:{train_loss.item():.4f} train_time:{approx_time:.0f}ms step_avg:{approx_time/timed_steps:.2f}ms")

if master_process:
    print(f"peak memory consumption: {torch.cuda.max_memory_allocated() // 1024 // 1024} MiB")

# -------------------------------------------------------------------------
# clean up nice
dist.destroy_process_group()
====================================================================================================
Running pytorch 2.6.0.dev20241203+cu124 compiled for CUDA 12.4
nvidia-smi:
Thu Dec  5 04:19:11 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.183.06             Driver Version: 535.183.06   CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA H100 80GB HBM3          On  | 00000000:19:00.0 Off |                    0 |
| N/A   38C    P0              75W / 700W |      3MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   1  NVIDIA H100 80GB HBM3          On  | 00000000:3B:00.0 Off |                    0 |
| N/A   30C    P0             115W / 700W |    529MiB / 81559MiB |      1%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   2  NVIDIA H100 80GB HBM3          On  | 00000000:4C:00.0 Off |                    0 |
| N/A   31C    P0             118W / 700W |    529MiB / 81559MiB |      1%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   3  NVIDIA H100 80GB HBM3          On  | 00000000:5D:00.0 Off |                    0 |
| N/A   37C    P0              96W / 700W |     22MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   4  NVIDIA H100 80GB HBM3          On  | 00000000:9B:00.0 Off |                    0 |
| N/A   38C    P0             122W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   5  NVIDIA H100 80GB HBM3          On  | 00000000:BB:00.0 Off |                    0 |
| N/A   29C    P0             110W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   6  NVIDIA H100 80GB HBM3          On  | 00000000:CB:00.0 Off |                    0 |
| N/A   38C    P0             127W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   7  NVIDIA H100 80GB HBM3          On  | 00000000:DB:00.0 Off |                    0 |
| N/A   29C    P0             119W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
+---------------------------------------------------------------------------------------+

====================================================================================================
Training DataLoader: total number of tokens: 1100000000 across 11 files
Validation DataLoader: total number of tokens: 100000000 across 1 files
====================================================================================================
step:0/1530 val_loss:10.8258 train_time:0ms step_avg:nanms
step:1/1530 train_loss:10.8258 train_time:31700ms step_avg:nanms
step:2/1530 train_loss:10.0712 train_time:31810ms step_avg:nanms
step:3/1530 train_loss:8.3189 train_time:31971ms step_avg:nanms
step:4/1530 train_loss:7.6580 train_time:32130ms step_avg:nanms
step:5/1530 train_loss:7.5266 train_time:32291ms step_avg:nanms
step:6/1530 train_loss:7.0569 train_time:32452ms step_avg:nanms
step:7/1530 train_loss:7.1731 train_time:32612ms step_avg:nanms
step:8/1530 train_loss:6.7722 train_time:32773ms step_avg:nanms
step:9/1530 train_loss:6.6751 train_time:32933ms step_avg:nanms
step:10/1530 train_loss:6.6186 train_time:33093ms step_avg:nanms
step:11/1530 train_loss:6.4493 train_time:115ms step_avg:nanms
step:12/1530 train_loss:6.3193 train_time:275ms step_avg:nanms
step:13/1530 train_loss:6.2083 train_time:435ms step_avg:145.11ms
step:14/1530 train_loss:6.1643 train_time:594ms step_avg:148.53ms
step:15/1530 train_loss:6.1454 train_time:755ms step_avg:151.00ms
step:16/1530 train_loss:6.1973 train_time:916ms step_avg:152.65ms
step:17/1530 train_loss:6.1793 train_time:1076ms step_avg:153.71ms
step:18/1530 train_loss:5.9431 train_time:1237ms step_avg:154.61ms
step:19/1530 train_loss:5.9657 train_time:1397ms step_avg:155.20ms
step:20/1530 train_loss:5.6992 train_time:1558ms step_avg:155.78ms
step:21/1530 train_loss:5.9628 train_time:1719ms step_avg:156.23ms
step:22/1530 train_loss:6.1842 train_time:1880ms step_avg:156.65ms
step:23/1530 train_loss:5.8591 train_time:2040ms step_avg:156.91ms
step:24/1530 train_loss:6.0422 train_time:2201ms step_avg:157.18ms
step:25/1530 train_loss:5.6847 train_time:2362ms step_avg:157.46ms
step:26/1530 train_loss:5.5936 train_time:2523ms step_avg:157.69ms
step:27/1530 train_loss:5.7594 train_time:2683ms step_avg:157.82ms
step:28/1530 train_loss:5.4221 train_time:2844ms step_avg:158.00ms
step:29/1530 train_loss:5.6559 train_time:3003ms step_avg:158.08ms
step:30/1530 train_loss:5.4636 train_time:3163ms step_avg:158.14ms
step:31/1530 train_loss:5.4368 train_time:3324ms step_avg:158.27ms
step:32/1530 train_loss:5.2958 train_time:3484ms step_avg:158.35ms
step:33/1530 train_loss:5.5847 train_time:3645ms step_avg:158.48ms
step:34/1530 train_loss:5.5105 train_time:3805ms step_avg:158.55ms
step:35/1530 train_loss:5.6142 train_time:3966ms step_avg:158.62ms
step:36/1530 train_loss:5.5516 train_time:4127ms step_avg:158.73ms
step:37/1530 train_loss:5.4632 train_time:4287ms step_avg:158.77ms
step:38/1530 train_loss:5.3200 train_time:4449ms step_avg:158.90ms
step:39/1530 train_loss:5.3236 train_time:4610ms step_avg:158.96ms
step:40/1530 train_loss:5.2617 train_time:4770ms step_avg:159.01ms
step:41/1530 train_loss:5.2323 train_time:4931ms step_avg:159.07ms
step:42/1530 train_loss:5.1701 train_time:5092ms step_avg:159.11ms
step:43/1530 train_loss:5.2653 train_time:5252ms step_avg:159.17ms
step:44/1530 train_loss:5.2457 train_time:5413ms step_avg:159.20ms
step:45/1530 train_loss:5.3762 train_time:5573ms step_avg:159.23ms
step:46/1530 train_loss:5.1711 train_time:5734ms step_avg:159.28ms
step:47/1530 train_loss:5.0591 train_time:5894ms step_avg:159.29ms
step:48/1530 train_loss:5.1994 train_time:6055ms step_avg:159.33ms
step:49/1530 train_loss:5.1457 train_time:6214ms step_avg:159.34ms
step:50/1530 train_loss:5.2598 train_time:6375ms step_avg:159.38ms
step:51/1530 train_loss:5.1425 train_time:6536ms step_avg:159.41ms
step:52/1530 train_loss:5.0354 train_time:6695ms step_avg:159.41ms
step:53/1530 train_loss:5.1631 train_time:6856ms step_avg:159.44ms
step:54/1530 train_loss:5.0209 train_time:7016ms step_avg:159.46ms
step:55/1530 train_loss:5.4221 train_time:7177ms step_avg:159.48ms
step:56/1530 train_loss:5.0301 train_time:7338ms step_avg:159.51ms
step:57/1530 train_loss:4.8777 train_time:7497ms step_avg:159.52ms
step:58/1530 train_loss:5.0385 train_time:7658ms step_avg:159.54ms
step:59/1530 train_loss:5.0233 train_time:7818ms step_avg:159.55ms
step:60/1530 train_loss:5.1404 train_time:7978ms step_avg:159.56ms
step:61/1530 train_loss:4.8576 train_time:8138ms step_avg:159.58ms
step:62/1530 train_loss:4.9707 train_time:8299ms step_avg:159.59ms
step:63/1530 train_loss:4.9647 train_time:8458ms step_avg:159.59ms
step:64/1530 train_loss:4.9741 train_time:8619ms step_avg:159.61ms
step:65/1530 train_loss:4.8013 train_time:8779ms step_avg:159.62ms
step:66/1530 train_loss:4.9176 train_time:8939ms step_avg:159.63ms
step:67/1530 train_loss:4.8238 train_time:9099ms step_avg:159.63ms
step:68/1530 train_loss:5.1131 train_time:9259ms step_avg:159.63ms
step:69/1530 train_loss:4.7514 train_time:9420ms step_avg:159.65ms
step:70/1530 train_loss:4.8503 train_time:9579ms step_avg:159.65ms
step:71/1530 train_loss:4.9771 train_time:9739ms step_avg:159.66ms
step:72/1530 train_loss:4.8860 train_time:9899ms step_avg:159.67ms
step:73/1530 train_loss:4.7690 train_time:10060ms step_avg:159.68ms
step:74/1530 train_loss:4.9002 train_time:10220ms step_avg:159.68ms
step:75/1530 train_loss:4.8690 train_time:10380ms step_avg:159.69ms
step:76/1530 train_loss:4.7941 train_time:10541ms step_avg:159.71ms
step:77/1530 train_loss:4.9041 train_time:10700ms step_avg:159.71ms
step:78/1530 train_loss:5.0988 train_time:10861ms step_avg:159.72ms
step:79/1530 train_loss:4.8210 train_time:11022ms step_avg:159.74ms
step:80/1530 train_loss:4.8638 train_time:11183ms step_avg:159.76ms
step:81/1530 train_loss:4.6510 train_time:11344ms step_avg:159.78ms
step:82/1530 train_loss:4.8158 train_time:11504ms step_avg:159.78ms
step:83/1530 train_loss:4.7751 train_time:11665ms step_avg:159.79ms
step:84/1530 train_loss:4.7739 train_time:11825ms step_avg:159.80ms
step:85/1530 train_loss:4.6251 train_time:11984ms step_avg:159.79ms
step:86/1530 train_loss:4.8415 train_time:12146ms step_avg:159.82ms
step:87/1530 train_loss:4.7513 train_time:12307ms step_avg:159.83ms
step:88/1530 train_loss:4.7527 train_time:12467ms step_avg:159.83ms
step:89/1530 train_loss:4.7185 train_time:12627ms step_avg:159.84ms
step:90/1530 train_loss:4.6583 train_time:12788ms step_avg:159.85ms
step:91/1530 train_loss:4.6366 train_time:12950ms step_avg:159.88ms
step:92/1530 train_loss:4.7982 train_time:13111ms step_avg:159.88ms
step:93/1530 train_loss:4.6244 train_time:13272ms step_avg:159.90ms
step:94/1530 train_loss:4.6435 train_time:13432ms step_avg:159.91ms
step:95/1530 train_loss:4.6931 train_time:13592ms step_avg:159.91ms
step:96/1530 train_loss:4.5841 train_time:13754ms step_avg:159.93ms
step:97/1530 train_loss:4.6452 train_time:13914ms step_avg:159.93ms
step:98/1530 train_loss:4.5891 train_time:14074ms step_avg:159.93ms
step:99/1530 train_loss:4.6667 train_time:14235ms step_avg:159.94ms
step:100/1530 train_loss:4.6730 train_time:14395ms step_avg:159.94ms
step:101/1530 train_loss:4.5283 train_time:14556ms step_avg:159.95ms
step:102/1530 train_loss:4.7004 train_time:14716ms step_avg:159.95ms
step:103/1530 train_loss:4.5829 train_time:14876ms step_avg:159.96ms
step:104/1530 train_loss:4.5374 train_time:15038ms step_avg:159.98ms
step:105/1530 train_loss:4.5638 train_time:15197ms step_avg:159.97ms
step:106/1530 train_loss:4.6033 train_time:15358ms step_avg:159.98ms
step:107/1530 train_loss:4.4989 train_time:15518ms step_avg:159.98ms
step:108/1530 train_loss:4.3547 train_time:15678ms step_avg:159.98ms
step:109/1530 train_loss:4.4911 train_time:15839ms step_avg:159.99ms
step:110/1530 train_loss:4.5133 train_time:15999ms step_avg:159.99ms
step:111/1530 train_loss:4.4305 train_time:16159ms step_avg:159.99ms
step:112/1530 train_loss:4.5893 train_time:16320ms step_avg:160.00ms
step:113/1530 train_loss:4.5020 train_time:16481ms step_avg:160.01ms
step:114/1530 train_loss:4.3681 train_time:16642ms step_avg:160.02ms
step:115/1530 train_loss:4.5105 train_time:16804ms step_avg:160.04ms
step:116/1530 train_loss:4.4765 train_time:16968ms step_avg:160.08ms
step:117/1530 train_loss:4.3820 train_time:17133ms step_avg:160.12ms
step:118/1530 train_loss:4.5929 train_time:17296ms step_avg:160.14ms
step:119/1530 train_loss:4.4578 train_time:17460ms step_avg:160.18ms
step:120/1530 train_loss:4.3308 train_time:17625ms step_avg:160.22ms
step:121/1530 train_loss:4.3003 train_time:17789ms step_avg:160.26ms
step:122/1530 train_loss:4.4568 train_time:17954ms step_avg:160.31ms
step:123/1530 train_loss:4.2898 train_time:18117ms step_avg:160.33ms
step:124/1530 train_loss:4.5889 train_time:18280ms step_avg:160.36ms
step:125/1530 train_loss:4.4705 train_time:18445ms step_avg:160.39ms
step:125/1530 val_loss:4.4100 train_time:18492ms step_avg:160.80ms
step:126/1530 train_loss:4.4242 train_time:18613ms step_avg:160.45ms
step:127/1530 train_loss:4.4361 train_time:18777ms step_avg:160.48ms
step:128/1530 train_loss:4.3753 train_time:18941ms step_avg:160.51ms
step:129/1530 train_loss:4.6770 train_time:19104ms step_avg:160.54ms
step:130/1530 train_loss:4.3588 train_time:19268ms step_avg:160.57ms
step:131/1530 train_loss:4.3871 train_time:19433ms step_avg:160.60ms
step:132/1530 train_loss:4.3478 train_time:19596ms step_avg:160.62ms
step:133/1530 train_loss:4.4543 train_time:19759ms step_avg:160.64ms
step:134/1530 train_loss:4.2747 train_time:19924ms step_avg:160.67ms
step:135/1530 train_loss:4.4699 train_time:20089ms step_avg:160.71ms
step:136/1530 train_loss:4.2281 train_time:20253ms step_avg:160.74ms
step:137/1530 train_loss:4.3742 train_time:20416ms step_avg:160.76ms
step:138/1530 train_loss:4.2946 train_time:20580ms step_avg:160.78ms
step:139/1530 train_loss:4.3925 train_time:20744ms step_avg:160.81ms
step:140/1530 train_loss:4.4700 train_time:20910ms step_avg:160.84ms
step:141/1530 train_loss:4.3150 train_time:21073ms step_avg:160.86ms
step:142/1530 train_loss:4.3124 train_time:21237ms step_avg:160.88ms
step:143/1530 train_loss:4.2564 train_time:21401ms step_avg:160.91ms
step:144/1530 train_loss:4.3530 train_time:21566ms step_avg:160.94ms
step:145/1530 train_loss:4.3077 train_time:21731ms step_avg:160.97ms
step:146/1530 train_loss:4.1713 train_time:21894ms step_avg:160.99ms
step:147/1530 train_loss:4.3229 train_time:22058ms step_avg:161.01ms
step:148/1530 train_loss:4.3511 train_time:22222ms step_avg:161.03ms
step:149/1530 train_loss:4.2912 train_time:22386ms step_avg:161.05ms
step:150/1530 train_loss:4.4345 train_time:22550ms step_avg:161.07ms
step:151/1530 train_loss:4.2782 train_time:22715ms step_avg:161.10ms
step:152/1530 train_loss:4.2828 train_time:22878ms step_avg:161.11ms
step:153/1530 train_loss:4.3597 train_time:23041ms step_avg:161.12ms
step:154/1530 train_loss:4.3660 train_time:23205ms step_avg:161.15ms
step:155/1530 train_loss:4.2736 train_time:23370ms step_avg:161.17ms
step:156/1530 train_loss:4.3512 train_time:23533ms step_avg:161.19ms
step:157/1530 train_loss:4.4102 train_time:23696ms step_avg:161.20ms
step:158/1530 train_loss:4.2496 train_time:23859ms step_avg:161.21ms
step:159/1530 train_loss:4.3079 train_time:24024ms step_avg:161.24ms
step:160/1530 train_loss:4.1286 train_time:24189ms step_avg:161.26ms
step:161/1530 train_loss:4.3402 train_time:24352ms step_avg:161.27ms
step:162/1530 train_loss:4.3522 train_time:24516ms step_avg:161.29ms
step:163/1530 train_loss:4.3343 train_time:24679ms step_avg:161.30ms
step:164/1530 train_loss:4.1767 train_time:24843ms step_avg:161.32ms
step:165/1530 train_loss:4.2865 train_time:25008ms step_avg:161.34ms
step:166/1530 train_loss:4.3334 train_time:25171ms step_avg:161.35ms
step:167/1530 train_loss:4.1928 train_time:25334ms step_avg:161.36ms
step:168/1530 train_loss:4.2739 train_time:25497ms step_avg:161.38ms
step:169/1530 train_loss:4.1495 train_time:25662ms step_avg:161.39ms
step:170/1530 train_loss:4.0222 train_time:25829ms step_avg:161.43ms
step:171/1530 train_loss:4.1979 train_time:25993ms step_avg:161.45ms
step:172/1530 train_loss:4.2049 train_time:26155ms step_avg:161.45ms
step:173/1530 train_loss:4.2582 train_time:26319ms step_avg:161.46ms
step:174/1530 train_loss:4.4201 train_time:26482ms step_avg:161.47ms
step:175/1530 train_loss:4.2362 train_time:26644ms step_avg:161.48ms
step:176/1530 train_loss:4.0884 train_time:26806ms step_avg:161.48ms
step:177/1530 train_loss:4.0647 train_time:26969ms step_avg:161.49ms
step:178/1530 train_loss:4.1805 train_time:27132ms step_avg:161.50ms
step:179/1530 train_loss:4.1231 train_time:27295ms step_avg:161.51ms
step:180/1530 train_loss:4.1086 train_time:27457ms step_avg:161.51ms
step:181/1530 train_loss:4.2923 train_time:27620ms step_avg:161.52ms
step:182/1530 train_loss:4.1428 train_time:27782ms step_avg:161.52ms
step:183/1530 train_loss:4.1197 train_time:27945ms step_avg:161.53ms
step:184/1530 train_loss:4.1269 train_time:28107ms step_avg:161.54ms
step:185/1530 train_loss:4.2000 train_time:28269ms step_avg:161.54ms
step:186/1530 train_loss:4.1650 train_time:28433ms step_avg:161.55ms
step:187/1530 train_loss:4.2386 train_time:28595ms step_avg:161.55ms
step:188/1530 train_loss:4.1664 train_time:28892ms step_avg:162.32ms
step:189/1530 train_loss:4.1125 train_time:29224ms step_avg:163.26ms
step:190/1530 train_loss:4.2029 train_time:29386ms step_avg:163.26ms
step:191/1530 train_loss:4.0841 train_time:29550ms step_avg:163.26ms
step:192/1530 train_loss:4.0201 train_time:29712ms step_avg:163.26ms
step:193/1530 train_loss:4.2478 train_time:29874ms step_avg:163.25ms
step:194/1530 train_loss:4.1649 train_time:30037ms step_avg:163.24ms
step:195/1530 train_loss:4.3449 train_time:30200ms step_avg:163.24ms
step:196/1530 train_loss:4.1703 train_time:30363ms step_avg:163.24ms
step:197/1530 train_loss:4.0353 train_time:30527ms step_avg:163.25ms
step:198/1530 train_loss:4.1732 train_time:30690ms step_avg:163.24ms
step:199/1530 train_loss:4.0237 train_time:30852ms step_avg:163.24ms
step:200/1530 train_loss:4.1119 train_time:31014ms step_avg:163.23ms
step:201/1530 train_loss:3.9941 train_time:31177ms step_avg:163.23ms
step:202/1530 train_loss:4.2469 train_time:31339ms step_avg:163.23ms
step:203/1530 train_loss:4.0618 train_time:31504ms step_avg:163.23ms
step:204/1530 train_loss:4.1826 train_time:31667ms step_avg:163.23ms
step:205/1530 train_loss:4.2389 train_time:31831ms step_avg:163.23ms
step:206/1530 train_loss:3.9401 train_time:31994ms step_avg:163.23ms
step:207/1530 train_loss:4.0718 train_time:32156ms step_avg:163.23ms
step:208/1530 train_loss:4.1032 train_time:32318ms step_avg:163.22ms
step:209/1530 train_loss:4.2417 train_time:32481ms step_avg:163.22ms
step:210/1530 train_loss:4.1769 train_time:32645ms step_avg:163.22ms
step:211/1530 train_loss:4.0573 train_time:32808ms step_avg:163.23ms
step:212/1530 train_loss:4.1186 train_time:32971ms step_avg:163.22ms
step:213/1530 train_loss:4.0422 train_time:33133ms step_avg:163.22ms
step:214/1530 train_loss:4.1053 train_time:33295ms step_avg:163.21ms
step:215/1530 train_loss:3.9460 train_time:33457ms step_avg:163.20ms
step:216/1530 train_loss:3.9967 train_time:33620ms step_avg:163.21ms
step:217/1530 train_loss:3.9996 train_time:33783ms step_avg:163.20ms
step:218/1530 train_loss:4.0773 train_time:33947ms step_avg:163.21ms
step:219/1530 train_loss:4.0617 train_time:34111ms step_avg:163.21ms
step:220/1530 train_loss:4.0785 train_time:34273ms step_avg:163.20ms
step:221/1530 train_loss:4.0907 train_time:34435ms step_avg:163.20ms
step:222/1530 train_loss:3.9920 train_time:34598ms step_avg:163.20ms
step:223/1530 train_loss:3.9823 train_time:34760ms step_avg:163.19ms
step:224/1530 train_loss:4.2900 train_time:34922ms step_avg:163.19ms
step:225/1530 train_loss:3.9151 train_time:35086ms step_avg:163.19ms
step:226/1530 train_loss:3.9880 train_time:35251ms step_avg:163.20ms
step:227/1530 train_loss:3.9847 train_time:35413ms step_avg:163.19ms
step:228/1530 train_loss:4.1318 train_time:35577ms step_avg:163.20ms
step:229/1530 train_loss:3.9117 train_time:35742ms step_avg:163.21ms
step:230/1530 train_loss:4.0384 train_time:35908ms step_avg:163.22ms
step:231/1530 train_loss:3.8976 train_time:36073ms step_avg:163.23ms
step:232/1530 train_loss:3.9657 train_time:36239ms step_avg:163.24ms
step:233/1530 train_loss:4.0811 train_time:36406ms step_avg:163.25ms
step:234/1530 train_loss:4.0195 train_time:36571ms step_avg:163.26ms
step:235/1530 train_loss:3.8993 train_time:36738ms step_avg:163.28ms
step:236/1530 train_loss:4.0823 train_time:36903ms step_avg:163.29ms
step:237/1530 train_loss:4.0659 train_time:37069ms step_avg:163.30ms
step:238/1530 train_loss:3.9388 train_time:37235ms step_avg:163.31ms
step:239/1530 train_loss:4.0770 train_time:37401ms step_avg:163.32ms
step:240/1530 train_loss:4.1066 train_time:37568ms step_avg:163.34ms
step:241/1530 train_loss:3.9592 train_time:37733ms step_avg:163.35ms
step:242/1530 train_loss:4.1279 train_time:37899ms step_avg:163.36ms
step:243/1530 train_loss:4.0000 train_time:38065ms step_avg:163.37ms
step:244/1530 train_loss:4.0752 train_time:38232ms step_avg:163.39ms
step:245/1530 train_loss:4.1367 train_time:38397ms step_avg:163.39ms
step:246/1530 train_loss:4.0480 train_time:38562ms step_avg:163.40ms
step:247/1530 train_loss:3.9930 train_time:38731ms step_avg:163.42ms
step:248/1530 train_loss:4.0896 train_time:38896ms step_avg:163.43ms
step:249/1530 train_loss:3.9079 train_time:39062ms step_avg:163.44ms
step:250/1530 train_loss:3.9734 train_time:39230ms step_avg:163.46ms
step:250/1530 val_loss:3.9996 train_time:39277ms step_avg:163.65ms
step:251/1530 train_loss:4.0678 train_time:39397ms step_avg:163.47ms
step:252/1530 train_loss:4.1550 train_time:39563ms step_avg:163.49ms
step:253/1530 train_loss:3.9228 train_time:39731ms step_avg:163.50ms
step:254/1530 train_loss:3.8780 train_time:39898ms step_avg:163.51ms
step:255/1530 train_loss:4.0721 train_time:40063ms step_avg:163.52ms
step:256/1530 train_loss:3.9778 train_time:40230ms step_avg:163.53ms
step:257/1530 train_loss:3.9846 train_time:40396ms step_avg:163.54ms
step:258/1530 train_loss:3.9803 train_time:40561ms step_avg:163.55ms
step:259/1530 train_loss:4.0248 train_time:40728ms step_avg:163.57ms
step:260/1530 train_loss:4.0476 train_time:40896ms step_avg:163.58ms
step:261/1530 train_loss:4.0117 train_time:41062ms step_avg:163.59ms
step:262/1530 train_loss:3.9839 train_time:41228ms step_avg:163.60ms
step:263/1530 train_loss:3.8863 train_time:41395ms step_avg:163.62ms
step:264/1530 train_loss:3.9782 train_time:41561ms step_avg:163.63ms
step:265/1530 train_loss:3.8637 train_time:41728ms step_avg:163.64ms
step:266/1530 train_loss:3.9132 train_time:41895ms step_avg:163.65ms
step:267/1530 train_loss:3.9240 train_time:42060ms step_avg:163.66ms
step:268/1530 train_loss:3.9580 train_time:42225ms step_avg:163.66ms
step:269/1530 train_loss:3.8474 train_time:42393ms step_avg:163.68ms
step:270/1530 train_loss:4.0944 train_time:42559ms step_avg:163.69ms
step:271/1530 train_loss:3.9615 train_time:42725ms step_avg:163.70ms
step:272/1530 train_loss:3.9239 train_time:42892ms step_avg:163.71ms
step:273/1530 train_loss:3.9439 train_time:43058ms step_avg:163.72ms
step:274/1530 train_loss:4.0373 train_time:43224ms step_avg:163.73ms
step:275/1530 train_loss:4.0559 train_time:43391ms step_avg:163.74ms
step:276/1530 train_loss:4.2244 train_time:43556ms step_avg:163.75ms
step:277/1530 train_loss:4.0357 train_time:43721ms step_avg:163.75ms
step:278/1530 train_loss:4.0788 train_time:43888ms step_avg:163.76ms
step:279/1530 train_loss:3.9966 train_time:44055ms step_avg:163.77ms
step:280/1530 train_loss:4.1757 train_time:44222ms step_avg:163.78ms
step:281/1530 train_loss:3.9674 train_time:44387ms step_avg:163.79ms
step:282/1530 train_loss:3.9371 train_time:44555ms step_avg:163.81ms
step:283/1530 train_loss:3.9038 train_time:44721ms step_avg:163.81ms
step:284/1530 train_loss:4.0460 train_time:44887ms step_avg:163.82ms
step:285/1530 train_loss:4.0565 train_time:45053ms step_avg:163.83ms
step:286/1530 train_loss:4.0828 train_time:45218ms step_avg:163.83ms
step:287/1530 train_loss:3.8960 train_time:45384ms step_avg:163.84ms
step:288/1530 train_loss:3.9946 train_time:45549ms step_avg:163.85ms
step:289/1530 train_loss:3.8671 train_time:45715ms step_avg:163.85ms
step:290/1530 train_loss:3.8543 train_time:45880ms step_avg:163.86ms
step:291/1530 train_loss:3.8980 train_time:46045ms step_avg:163.86ms
step:292/1530 train_loss:3.8553 train_time:46210ms step_avg:163.86ms
step:293/1530 train_loss:3.8918 train_time:46376ms step_avg:163.87ms
step:294/1530 train_loss:3.9238 train_time:46539ms step_avg:163.87ms
step:295/1530 train_loss:3.8285 train_time:46704ms step_avg:163.87ms
step:296/1530 train_loss:3.8540 train_time:46871ms step_avg:163.88ms
step:297/1530 train_loss:3.8610 train_time:47035ms step_avg:163.89ms
step:298/1530 train_loss:3.9601 train_time:47200ms step_avg:163.89ms
step:299/1530 train_loss:3.8114 train_time:47365ms step_avg:163.89ms
step:300/1530 train_loss:3.9551 train_time:47530ms step_avg:163.90ms
step:301/1530 train_loss:3.9528 train_time:47695ms step_avg:163.90ms
step:302/1530 train_loss:3.9313 train_time:47860ms step_avg:163.90ms
step:303/1530 train_loss:3.9751 train_time:48025ms step_avg:163.91ms
step:304/1530 train_loss:3.9598 train_time:48191ms step_avg:163.92ms
step:305/1530 train_loss:4.4420 train_time:48356ms step_avg:163.92ms
step:306/1530 train_loss:3.9281 train_time:48520ms step_avg:163.92ms
step:307/1530 train_loss:3.8250 train_time:48687ms step_avg:163.93ms
step:308/1530 train_loss:3.9721 train_time:48853ms step_avg:163.94ms
step:309/1530 train_loss:3.8597 train_time:49017ms step_avg:163.94ms
step:310/1530 train_loss:4.0821 train_time:49183ms step_avg:163.94ms
step:311/1530 train_loss:3.9276 train_time:49349ms step_avg:163.95ms
step:312/1530 train_loss:3.8568 train_time:49513ms step_avg:163.95ms
step:313/1530 train_loss:3.9219 train_time:49679ms step_avg:163.96ms
step:314/1530 train_loss:4.0488 train_time:49844ms step_avg:163.96ms
step:315/1530 train_loss:3.9295 train_time:50011ms step_avg:163.97ms
step:316/1530 train_loss:3.7905 train_time:50176ms step_avg:163.97ms
step:317/1530 train_loss:3.8679 train_time:50340ms step_avg:163.97ms
step:318/1530 train_loss:3.9165 train_time:50506ms step_avg:163.98ms
step:319/1530 train_loss:3.8837 train_time:50672ms step_avg:163.99ms
step:320/1530 train_loss:4.0089 train_time:50836ms step_avg:163.99ms
step:321/1530 train_loss:3.9545 train_time:51001ms step_avg:163.99ms
step:322/1530 train_loss:3.9248 train_time:51167ms step_avg:164.00ms
step:323/1530 train_loss:3.9915 train_time:51332ms step_avg:164.00ms
step:324/1530 train_loss:3.9328 train_time:51497ms step_avg:164.00ms
step:325/1530 train_loss:4.0034 train_time:51662ms step_avg:164.01ms
step:326/1530 train_loss:3.8883 train_time:51828ms step_avg:164.01ms
step:327/1530 train_loss:4.3804 train_time:51994ms step_avg:164.02ms
step:328/1530 train_loss:4.0663 train_time:52159ms step_avg:164.02ms
step:329/1530 train_loss:3.7793 train_time:52325ms step_avg:164.03ms
step:330/1530 train_loss:3.7365 train_time:52492ms step_avg:164.04ms
step:331/1530 train_loss:3.9679 train_time:52657ms step_avg:164.04ms
step:332/1530 train_loss:3.9066 train_time:52822ms step_avg:164.04ms
step:333/1530 train_loss:3.8753 train_time:52988ms step_avg:164.05ms
step:334/1530 train_loss:3.8307 train_time:53153ms step_avg:164.05ms
step:335/1530 train_loss:4.0021 train_time:53318ms step_avg:164.06ms
step:336/1530 train_loss:3.9516 train_time:53482ms step_avg:164.05ms
step:337/1530 train_loss:4.4034 train_time:53649ms step_avg:164.07ms
step:338/1530 train_loss:3.9269 train_time:53815ms step_avg:164.07ms
step:339/1530 train_loss:3.8615 train_time:53980ms step_avg:164.07ms
step:340/1530 train_loss:3.9232 train_time:54145ms step_avg:164.08ms
step:341/1530 train_loss:3.8440 train_time:54313ms step_avg:164.09ms
step:342/1530 train_loss:3.8034 train_time:54480ms step_avg:164.10ms
step:343/1530 train_loss:3.8287 train_time:54649ms step_avg:164.11ms
step:344/1530 train_loss:3.9919 train_time:54817ms step_avg:164.12ms
step:345/1530 train_loss:3.8016 train_time:54985ms step_avg:164.13ms
step:346/1530 train_loss:3.7628 train_time:55153ms step_avg:164.15ms
step:347/1530 train_loss:3.7886 train_time:55320ms step_avg:164.15ms
step:348/1530 train_loss:3.8522 train_time:55491ms step_avg:164.18ms
step:349/1530 train_loss:3.8281 train_time:55661ms step_avg:164.19ms
step:350/1530 train_loss:3.5651 train_time:55831ms step_avg:164.21ms
step:351/1530 train_loss:3.8203 train_time:55998ms step_avg:164.22ms
step:352/1530 train_loss:4.1759 train_time:56166ms step_avg:164.23ms
step:353/1530 train_loss:3.6599 train_time:56334ms step_avg:164.24ms
step:354/1530 train_loss:3.9174 train_time:56500ms step_avg:164.25ms
step:355/1530 train_loss:3.7780 train_time:56671ms step_avg:164.26ms
step:356/1530 train_loss:3.8720 train_time:56838ms step_avg:164.27ms
step:357/1530 train_loss:3.7521 train_time:57006ms step_avg:164.28ms
step:358/1530 train_loss:3.8543 train_time:57174ms step_avg:164.29ms
step:359/1530 train_loss:3.7675 train_time:57342ms step_avg:164.30ms
step:360/1530 train_loss:3.4137 train_time:57512ms step_avg:164.32ms
step:361/1530 train_loss:4.0114 train_time:57680ms step_avg:164.33ms
step:362/1530 train_loss:3.9080 train_time:57848ms step_avg:164.34ms
step:363/1530 train_loss:3.8393 train_time:58015ms step_avg:164.35ms
step:364/1530 train_loss:3.7371 train_time:58182ms step_avg:164.36ms
step:365/1530 train_loss:3.9055 train_time:58352ms step_avg:164.37ms
step:366/1530 train_loss:3.8578 train_time:58521ms step_avg:164.38ms
step:367/1530 train_loss:3.8464 train_time:58689ms step_avg:164.40ms
step:368/1530 train_loss:3.8428 train_time:58857ms step_avg:164.40ms
step:369/1530 train_loss:3.7435 train_time:59024ms step_avg:164.41ms
step:370/1530 train_loss:3.8721 train_time:59193ms step_avg:164.42ms
step:371/1530 train_loss:3.7240 train_time:59360ms step_avg:164.43ms
step:372/1530 train_loss:3.6890 train_time:59528ms step_avg:164.44ms
step:373/1530 train_loss:3.9094 train_time:59695ms step_avg:164.45ms
step:374/1530 train_loss:3.8228 train_time:59862ms step_avg:164.46ms
step:375/1530 train_loss:3.7942 train_time:60032ms step_avg:164.47ms
step:375/1530 val_loss:3.8199 train_time:60081ms step_avg:164.61ms
step:376/1530 train_loss:3.8602 train_time:60203ms step_avg:164.49ms
step:377/1530 train_loss:3.7813 train_time:60503ms step_avg:164.86ms
step:378/1530 train_loss:3.8431 train_time:60680ms step_avg:164.89ms
step:379/1530 train_loss:3.8638 train_time:61003ms step_avg:165.32ms
step:380/1530 train_loss:3.9493 train_time:61171ms step_avg:165.33ms
step:381/1530 train_loss:3.8410 train_time:61341ms step_avg:165.34ms
step:382/1530 train_loss:3.7982 train_time:61509ms step_avg:165.35ms
step:383/1530 train_loss:3.7926 train_time:61677ms step_avg:165.36ms
step:384/1530 train_loss:3.8666 train_time:61846ms step_avg:165.36ms
step:385/1530 train_loss:3.7867 train_time:62015ms step_avg:165.37ms
step:386/1530 train_loss:3.8882 train_time:62182ms step_avg:165.38ms
step:387/1530 train_loss:4.0469 train_time:62350ms step_avg:165.38ms
step:388/1530 train_loss:3.7857 train_time:62518ms step_avg:165.39ms
step:389/1530 train_loss:3.7864 train_time:62685ms step_avg:165.40ms
step:390/1530 train_loss:3.8827 train_time:62854ms step_avg:165.40ms
step:391/1530 train_loss:3.8010 train_time:63021ms step_avg:165.41ms
step:392/1530 train_loss:3.9149 train_time:63188ms step_avg:165.41ms
step:393/1530 train_loss:3.7590 train_time:63357ms step_avg:165.42ms
step:394/1530 train_loss:3.8792 train_time:63526ms step_avg:165.43ms
step:395/1530 train_loss:3.6233 train_time:63694ms step_avg:165.44ms
step:396/1530 train_loss:3.8330 train_time:63863ms step_avg:165.45ms
step:397/1530 train_loss:3.8500 train_time:64030ms step_avg:165.45ms
step:398/1530 train_loss:3.8822 train_time:64200ms step_avg:165.46ms
step:399/1530 train_loss:3.7599 train_time:64366ms step_avg:165.47ms
step:400/1530 train_loss:3.8207 train_time:64534ms step_avg:165.47ms
step:401/1530 train_loss:3.9015 train_time:64702ms step_avg:165.48ms
step:402/1530 train_loss:3.8362 train_time:64868ms step_avg:165.48ms
step:403/1530 train_loss:3.9511 train_time:65038ms step_avg:165.49ms
step:404/1530 train_loss:3.6700 train_time:65205ms step_avg:165.49ms
step:405/1530 train_loss:3.7777 train_time:65371ms step_avg:165.50ms
step:406/1530 train_loss:4.0844 train_time:65540ms step_avg:165.51ms
step:407/1530 train_loss:3.7704 train_time:65707ms step_avg:165.51ms
step:408/1530 train_loss:3.8123 train_time:65873ms step_avg:165.51ms
step:409/1530 train_loss:3.8446 train_time:66041ms step_avg:165.52ms
step:410/1530 train_loss:3.7456 train_time:66207ms step_avg:165.52ms
step:411/1530 train_loss:3.7552 train_time:66374ms step_avg:165.52ms
step:412/1530 train_loss:4.1606 train_time:66543ms step_avg:165.53ms
step:413/1530 train_loss:3.6267 train_time:66710ms step_avg:165.53ms
step:414/1530 train_loss:4.0073 train_time:66877ms step_avg:165.54ms
step:415/1530 train_loss:3.7489 train_time:67044ms step_avg:165.54ms
step:416/1530 train_loss:3.7616 train_time:67211ms step_avg:165.55ms
step:417/1530 train_loss:3.9562 train_time:67380ms step_avg:165.55ms
step:418/1530 train_loss:3.6797 train_time:67547ms step_avg:165.56ms
step:419/1530 train_loss:3.7982 train_time:67714ms step_avg:165.56ms
step:420/1530 train_loss:3.6958 train_time:67881ms step_avg:165.56ms
step:421/1530 train_loss:3.6441 train_time:68047ms step_avg:165.56ms
step:422/1530 train_loss:3.7784 train_time:68215ms step_avg:165.57ms
step:423/1530 train_loss:3.8742 train_time:68381ms step_avg:165.57ms
step:424/1530 train_loss:3.6121 train_time:68547ms step_avg:165.57ms
step:425/1530 train_loss:3.7822 train_time:68715ms step_avg:165.58ms
step:426/1530 train_loss:3.6404 train_time:68882ms step_avg:165.58ms
step:427/1530 train_loss:3.8754 train_time:69049ms step_avg:165.59ms
step:428/1530 train_loss:3.8030 train_time:69217ms step_avg:165.59ms
step:429/1530 train_loss:3.7523 train_time:69383ms step_avg:165.59ms
step:430/1530 train_loss:3.7015 train_time:69550ms step_avg:165.59ms
step:431/1530 train_loss:3.6233 train_time:69718ms step_avg:165.60ms
step:432/1530 train_loss:3.7555 train_time:69884ms step_avg:165.60ms
step:433/1530 train_loss:3.8111 train_time:70051ms step_avg:165.60ms
step:434/1530 train_loss:3.7695 train_time:70219ms step_avg:165.61ms
step:435/1530 train_loss:3.8031 train_time:70385ms step_avg:165.61ms
step:436/1530 train_loss:3.8268 train_time:70552ms step_avg:165.61ms
step:437/1530 train_loss:3.7059 train_time:70719ms step_avg:165.62ms
step:438/1530 train_loss:3.6962 train_time:70885ms step_avg:165.62ms
step:439/1530 train_loss:3.7070 train_time:71052ms step_avg:165.62ms
step:440/1530 train_loss:3.8816 train_time:71220ms step_avg:165.63ms
step:441/1530 train_loss:3.7473 train_time:71387ms step_avg:165.63ms
step:442/1530 train_loss:3.7369 train_time:71555ms step_avg:165.64ms
step:443/1530 train_loss:3.6130 train_time:71722ms step_avg:165.64ms
step:444/1530 train_loss:3.9158 train_time:71888ms step_avg:165.64ms
step:445/1530 train_loss:3.8400 train_time:72055ms step_avg:165.64ms
step:446/1530 train_loss:3.8268 train_time:72223ms step_avg:165.65ms
step:447/1530 train_loss:3.7468 train_time:72390ms step_avg:165.65ms
step:448/1530 train_loss:3.8492 train_time:72558ms step_avg:165.66ms
step:449/1530 train_loss:3.6874 train_time:72725ms step_avg:165.66ms
step:450/1530 train_loss:3.7096 train_time:72890ms step_avg:165.66ms
step:451/1530 train_loss:3.5795 train_time:73060ms step_avg:165.67ms
step:452/1530 train_loss:3.7017 train_time:73227ms step_avg:165.67ms
step:453/1530 train_loss:3.6648 train_time:73394ms step_avg:165.67ms
step:454/1530 train_loss:3.6294 train_time:73562ms step_avg:165.68ms
step:455/1530 train_loss:3.8345 train_time:73730ms step_avg:165.68ms
step:456/1530 train_loss:3.7199 train_time:73900ms step_avg:165.70ms
step:457/1530 train_loss:3.7745 train_time:74071ms step_avg:165.71ms
step:458/1530 train_loss:3.8165 train_time:74241ms step_avg:165.72ms
step:459/1530 train_loss:3.6276 train_time:74412ms step_avg:165.73ms
step:460/1530 train_loss:3.7853 train_time:74581ms step_avg:165.74ms
step:461/1530 train_loss:3.6834 train_time:74752ms step_avg:165.75ms
step:462/1530 train_loss:3.7240 train_time:74922ms step_avg:165.76ms
step:463/1530 train_loss:3.7689 train_time:75092ms step_avg:165.77ms
step:464/1530 train_loss:3.7060 train_time:75262ms step_avg:165.78ms
step:465/1530 train_loss:3.7083 train_time:75431ms step_avg:165.78ms
step:466/1530 train_loss:3.7922 train_time:75601ms step_avg:165.79ms
step:467/1530 train_loss:3.8124 train_time:75771ms step_avg:165.80ms
step:468/1530 train_loss:3.7866 train_time:75941ms step_avg:165.81ms
step:469/1530 train_loss:3.6786 train_time:76108ms step_avg:165.81ms
step:470/1530 train_loss:3.7558 train_time:76280ms step_avg:165.83ms
step:471/1530 train_loss:3.8000 train_time:76450ms step_avg:165.84ms
step:472/1530 train_loss:3.7781 train_time:76623ms step_avg:165.85ms
step:473/1530 train_loss:3.7112 train_time:76792ms step_avg:165.86ms
step:474/1530 train_loss:3.5921 train_time:76961ms step_avg:165.86ms
step:475/1530 train_loss:4.0169 train_time:77130ms step_avg:165.87ms
step:476/1530 train_loss:3.7440 train_time:77300ms step_avg:165.88ms
step:477/1530 train_loss:3.5931 train_time:77470ms step_avg:165.89ms
step:478/1530 train_loss:3.8193 train_time:77641ms step_avg:165.90ms
step:479/1530 train_loss:3.7619 train_time:77812ms step_avg:165.91ms
step:480/1530 train_loss:3.9207 train_time:77982ms step_avg:165.92ms
step:481/1530 train_loss:3.7192 train_time:78151ms step_avg:165.93ms
step:482/1530 train_loss:3.5206 train_time:78321ms step_avg:165.93ms
step:483/1530 train_loss:3.7997 train_time:78488ms step_avg:165.94ms
step:484/1530 train_loss:3.6517 train_time:78662ms step_avg:165.95ms
step:485/1530 train_loss:3.6472 train_time:78832ms step_avg:165.96ms
step:486/1530 train_loss:3.5632 train_time:79003ms step_avg:165.97ms
step:487/1530 train_loss:3.6754 train_time:79172ms step_avg:165.98ms
step:488/1530 train_loss:3.8734 train_time:79342ms step_avg:165.99ms
step:489/1530 train_loss:3.7045 train_time:79511ms step_avg:165.99ms
step:490/1530 train_loss:3.5862 train_time:79681ms step_avg:166.00ms
step:491/1530 train_loss:3.6049 train_time:79849ms step_avg:166.01ms
step:492/1530 train_loss:3.7280 train_time:80020ms step_avg:166.02ms
step:493/1530 train_loss:3.5689 train_time:80191ms step_avg:166.03ms
step:494/1530 train_loss:3.6965 train_time:80360ms step_avg:166.03ms
step:495/1530 train_loss:3.6606 train_time:80531ms step_avg:166.04ms
step:496/1530 train_loss:3.5023 train_time:80703ms step_avg:166.05ms
step:497/1530 train_loss:3.7311 train_time:80870ms step_avg:166.06ms
step:498/1530 train_loss:3.7746 train_time:81041ms step_avg:166.07ms
step:499/1530 train_loss:3.8106 train_time:81210ms step_avg:166.07ms
step:500/1530 train_loss:3.7253 train_time:81381ms step_avg:166.08ms
step:500/1530 val_loss:3.6985 train_time:81429ms step_avg:166.18ms
step:501/1530 train_loss:3.8042 train_time:81550ms step_avg:166.09ms
step:502/1530 train_loss:3.7448 train_time:81721ms step_avg:166.10ms
step:503/1530 train_loss:3.7696 train_time:81890ms step_avg:166.11ms
step:504/1530 train_loss:3.7128 train_time:82058ms step_avg:166.11ms
step:505/1530 train_loss:3.8020 train_time:82227ms step_avg:166.12ms
step:506/1530 train_loss:3.6553 train_time:82396ms step_avg:166.12ms
step:507/1530 train_loss:3.7624 train_time:82566ms step_avg:166.13ms
step:508/1530 train_loss:3.8181 train_time:82737ms step_avg:166.14ms
step:509/1530 train_loss:3.7659 train_time:82906ms step_avg:166.14ms
step:510/1530 train_loss:3.5763 train_time:83075ms step_avg:166.15ms
step:511/1530 train_loss:3.7706 train_time:83244ms step_avg:166.16ms
step:512/1530 train_loss:3.7184 train_time:83413ms step_avg:166.16ms
step:513/1530 train_loss:3.6646 train_time:83582ms step_avg:166.17ms
step:514/1530 train_loss:3.8057 train_time:83753ms step_avg:166.18ms
step:515/1530 train_loss:3.7307 train_time:83921ms step_avg:166.18ms
step:516/1530 train_loss:4.0691 train_time:84091ms step_avg:166.19ms
step:517/1530 train_loss:3.6772 train_time:84260ms step_avg:166.19ms
step:518/1530 train_loss:3.7635 train_time:84429ms step_avg:166.20ms
step:519/1530 train_loss:3.6474 train_time:84598ms step_avg:166.20ms
step:520/1530 train_loss:3.6737 train_time:84768ms step_avg:166.21ms
step:521/1530 train_loss:3.6589 train_time:84936ms step_avg:166.21ms
step:522/1530 train_loss:3.6618 train_time:85106ms step_avg:166.22ms
step:523/1530 train_loss:4.2799 train_time:85274ms step_avg:166.23ms
step:524/1530 train_loss:3.7259 train_time:85442ms step_avg:166.23ms
step:525/1530 train_loss:3.6773 train_time:85610ms step_avg:166.23ms
step:526/1530 train_loss:3.6912 train_time:85778ms step_avg:166.24ms
step:527/1530 train_loss:3.6544 train_time:85948ms step_avg:166.24ms
step:528/1530 train_loss:3.6207 train_time:86116ms step_avg:166.25ms
step:529/1530 train_loss:3.8454 train_time:86286ms step_avg:166.25ms
step:530/1530 train_loss:3.6434 train_time:86454ms step_avg:166.26ms
step:531/1530 train_loss:3.9130 train_time:86625ms step_avg:166.27ms
step:532/1530 train_loss:3.7219 train_time:86793ms step_avg:166.27ms
step:533/1530 train_loss:3.6461 train_time:86963ms step_avg:166.28ms
step:534/1530 train_loss:3.6626 train_time:87131ms step_avg:166.28ms
step:535/1530 train_loss:3.6001 train_time:87300ms step_avg:166.29ms
step:536/1530 train_loss:3.7473 train_time:87471ms step_avg:166.29ms
step:537/1530 train_loss:3.7165 train_time:87640ms step_avg:166.30ms
step:538/1530 train_loss:3.6190 train_time:87810ms step_avg:166.31ms
step:539/1530 train_loss:4.1094 train_time:87981ms step_avg:166.32ms
step:540/1530 train_loss:3.6683 train_time:88150ms step_avg:166.32ms
step:541/1530 train_loss:3.7804 train_time:88319ms step_avg:166.33ms
step:542/1530 train_loss:3.5864 train_time:88488ms step_avg:166.33ms
step:543/1530 train_loss:3.5841 train_time:88657ms step_avg:166.34ms
step:544/1530 train_loss:3.6345 train_time:88826ms step_avg:166.34ms
step:545/1530 train_loss:3.5855 train_time:88994ms step_avg:166.34ms
step:546/1530 train_loss:3.6158 train_time:89165ms step_avg:166.35ms
step:547/1530 train_loss:3.6244 train_time:89333ms step_avg:166.35ms
step:548/1530 train_loss:3.6086 train_time:89503ms step_avg:166.36ms
step:549/1530 train_loss:3.7141 train_time:89672ms step_avg:166.37ms
step:550/1530 train_loss:3.6151 train_time:89842ms step_avg:166.37ms
step:551/1530 train_loss:3.6228 train_time:90010ms step_avg:166.38ms
step:552/1530 train_loss:3.9219 train_time:90179ms step_avg:166.38ms
step:553/1530 train_loss:3.7527 train_time:90348ms step_avg:166.39ms
step:554/1530 train_loss:3.7050 train_time:90515ms step_avg:166.39ms
step:555/1530 train_loss:3.6225 train_time:90686ms step_avg:166.40ms
step:556/1530 train_loss:3.6957 train_time:90854ms step_avg:166.40ms
step:557/1530 train_loss:3.3072 train_time:91024ms step_avg:166.41ms
step:558/1530 train_loss:3.6109 train_time:91193ms step_avg:166.41ms
step:559/1530 train_loss:3.6471 train_time:91362ms step_avg:166.42ms
step:560/1530 train_loss:3.6818 train_time:91531ms step_avg:166.42ms
step:561/1530 train_loss:3.6093 train_time:91699ms step_avg:166.42ms
step:562/1530 train_loss:3.5512 train_time:91869ms step_avg:166.43ms
step:563/1530 train_loss:3.7557 train_time:92037ms step_avg:166.43ms
step:564/1530 train_loss:3.5665 train_time:92206ms step_avg:166.44ms
step:565/1530 train_loss:3.6780 train_time:92374ms step_avg:166.44ms
step:566/1530 train_loss:3.6045 train_time:92679ms step_avg:166.69ms
step:567/1530 train_loss:3.5988 train_time:92860ms step_avg:166.71ms
step:568/1530 train_loss:3.6787 train_time:93029ms step_avg:166.72ms
step:569/1530 train_loss:3.6378 train_time:93354ms step_avg:167.00ms
step:570/1530 train_loss:3.6862 train_time:93525ms step_avg:167.01ms
step:571/1530 train_loss:3.7540 train_time:93696ms step_avg:167.02ms
step:572/1530 train_loss:3.7243 train_time:93868ms step_avg:167.02ms
step:573/1530 train_loss:3.7339 train_time:94038ms step_avg:167.03ms
step:574/1530 train_loss:3.7755 train_time:94211ms step_avg:167.04ms
step:575/1530 train_loss:3.7220 train_time:94382ms step_avg:167.05ms
step:576/1530 train_loss:3.7513 train_time:94553ms step_avg:167.05ms
step:577/1530 train_loss:3.6601 train_time:94724ms step_avg:167.06ms
step:578/1530 train_loss:3.6672 train_time:94898ms step_avg:167.07ms
step:579/1530 train_loss:3.6643 train_time:95070ms step_avg:167.08ms
step:580/1530 train_loss:3.5838 train_time:95240ms step_avg:167.09ms
step:581/1530 train_loss:3.6319 train_time:95410ms step_avg:167.09ms
step:582/1530 train_loss:3.8407 train_time:95580ms step_avg:167.10ms
step:583/1530 train_loss:3.6224 train_time:95751ms step_avg:167.10ms
step:584/1530 train_loss:3.5873 train_time:95924ms step_avg:167.12ms
step:585/1530 train_loss:3.7824 train_time:96095ms step_avg:167.12ms
step:586/1530 train_loss:3.5082 train_time:96269ms step_avg:167.13ms
step:587/1530 train_loss:3.6601 train_time:96440ms step_avg:167.14ms
step:588/1530 train_loss:3.6359 train_time:96609ms step_avg:167.14ms
step:589/1530 train_loss:3.9918 train_time:96782ms step_avg:167.15ms
step:590/1530 train_loss:3.7822 train_time:96953ms step_avg:167.16ms
step:591/1530 train_loss:3.4969 train_time:97126ms step_avg:167.17ms
step:592/1530 train_loss:3.5269 train_time:97301ms step_avg:167.18ms
step:593/1530 train_loss:3.4965 train_time:97473ms step_avg:167.19ms
step:594/1530 train_loss:3.5449 train_time:97646ms step_avg:167.20ms
step:595/1530 train_loss:3.9071 train_time:97818ms step_avg:167.21ms
step:596/1530 train_loss:3.6400 train_time:97991ms step_avg:167.22ms
step:597/1530 train_loss:3.5814 train_time:98162ms step_avg:167.23ms
step:598/1530 train_loss:3.6526 train_time:98332ms step_avg:167.23ms
step:599/1530 train_loss:3.4732 train_time:98504ms step_avg:167.24ms
step:600/1530 train_loss:3.5932 train_time:98675ms step_avg:167.25ms
step:601/1530 train_loss:3.6425 train_time:98850ms step_avg:167.26ms
step:602/1530 train_loss:3.6665 train_time:99021ms step_avg:167.27ms
step:603/1530 train_loss:3.7768 train_time:99192ms step_avg:167.27ms
step:604/1530 train_loss:3.6006 train_time:99365ms step_avg:167.28ms
step:605/1530 train_loss:3.6113 train_time:99536ms step_avg:167.29ms
step:606/1530 train_loss:3.5712 train_time:99709ms step_avg:167.30ms
step:607/1530 train_loss:3.8309 train_time:99881ms step_avg:167.31ms
step:608/1530 train_loss:3.6292 train_time:100053ms step_avg:167.31ms
step:609/1530 train_loss:3.6143 train_time:100225ms step_avg:167.32ms
step:610/1530 train_loss:3.6982 train_time:100395ms step_avg:167.32ms
step:611/1530 train_loss:3.5984 train_time:100567ms step_avg:167.33ms
step:612/1530 train_loss:3.5673 train_time:100737ms step_avg:167.34ms
step:613/1530 train_loss:3.7528 train_time:100909ms step_avg:167.34ms
step:614/1530 train_loss:3.6934 train_time:101080ms step_avg:167.35ms
step:615/1530 train_loss:3.6883 train_time:101250ms step_avg:167.35ms
step:616/1530 train_loss:3.6257 train_time:101420ms step_avg:167.36ms
step:617/1530 train_loss:3.5474 train_time:101591ms step_avg:167.37ms
step:618/1530 train_loss:3.6843 train_time:101762ms step_avg:167.37ms
step:619/1530 train_loss:3.5436 train_time:101932ms step_avg:167.38ms
step:620/1530 train_loss:3.5833 train_time:102104ms step_avg:167.38ms
step:621/1530 train_loss:3.9205 train_time:102275ms step_avg:167.39ms
step:622/1530 train_loss:3.5642 train_time:102448ms step_avg:167.40ms
step:623/1530 train_loss:3.5929 train_time:102621ms step_avg:167.41ms
step:624/1530 train_loss:3.6859 train_time:102792ms step_avg:167.41ms
step:625/1530 train_loss:3.6977 train_time:102963ms step_avg:167.42ms
step:625/1530 val_loss:3.6175 train_time:103012ms step_avg:167.50ms
step:626/1530 train_loss:3.7349 train_time:103135ms step_avg:167.43ms
step:627/1530 train_loss:3.7122 train_time:103306ms step_avg:167.43ms
step:628/1530 train_loss:3.7562 train_time:103476ms step_avg:167.44ms
step:629/1530 train_loss:3.5882 train_time:103646ms step_avg:167.44ms
step:630/1530 train_loss:3.7127 train_time:103817ms step_avg:167.45ms
step:631/1530 train_loss:3.7355 train_time:103988ms step_avg:167.45ms
step:632/1530 train_loss:3.6412 train_time:104158ms step_avg:167.46ms
step:633/1530 train_loss:3.6026 train_time:104331ms step_avg:167.47ms
step:634/1530 train_loss:3.6955 train_time:104501ms step_avg:167.47ms
step:635/1530 train_loss:3.9443 train_time:104672ms step_avg:167.48ms
step:636/1530 train_loss:3.5400 train_time:104843ms step_avg:167.48ms
step:637/1530 train_loss:3.3481 train_time:105015ms step_avg:167.49ms
step:638/1530 train_loss:3.5927 train_time:105184ms step_avg:167.49ms
step:639/1530 train_loss:3.6363 train_time:105355ms step_avg:167.50ms
step:640/1530 train_loss:3.5659 train_time:105525ms step_avg:167.50ms
step:641/1530 train_loss:3.5828 train_time:105694ms step_avg:167.50ms
step:642/1530 train_loss:3.6259 train_time:105865ms step_avg:167.51ms
step:643/1530 train_loss:3.5887 train_time:106036ms step_avg:167.51ms
step:644/1530 train_loss:3.5593 train_time:106205ms step_avg:167.52ms
step:645/1530 train_loss:3.7689 train_time:106376ms step_avg:167.52ms
step:646/1530 train_loss:3.6652 train_time:106548ms step_avg:167.53ms
step:647/1530 train_loss:3.6539 train_time:106718ms step_avg:167.53ms
step:648/1530 train_loss:3.7033 train_time:106891ms step_avg:167.54ms
step:649/1530 train_loss:3.7576 train_time:107060ms step_avg:167.54ms
step:650/1530 train_loss:3.6151 train_time:107231ms step_avg:167.55ms
step:651/1530 train_loss:3.7653 train_time:107403ms step_avg:167.55ms
step:652/1530 train_loss:3.5813 train_time:107574ms step_avg:167.56ms
step:653/1530 train_loss:3.6600 train_time:107743ms step_avg:167.56ms
step:654/1530 train_loss:3.4263 train_time:107915ms step_avg:167.57ms
step:655/1530 train_loss:3.5793 train_time:108084ms step_avg:167.57ms
step:656/1530 train_loss:3.5730 train_time:108254ms step_avg:167.58ms
step:657/1530 train_loss:3.4925 train_time:108423ms step_avg:167.58ms
step:658/1530 train_loss:3.6800 train_time:108596ms step_avg:167.59ms
step:659/1530 train_loss:3.5796 train_time:108765ms step_avg:167.59ms
step:660/1530 train_loss:3.6756 train_time:108935ms step_avg:167.59ms
step:661/1530 train_loss:3.7473 train_time:109106ms step_avg:167.60ms
step:662/1530 train_loss:3.6584 train_time:109276ms step_avg:167.60ms
step:663/1530 train_loss:3.5453 train_time:109446ms step_avg:167.60ms
step:664/1530 train_loss:3.6082 train_time:109616ms step_avg:167.61ms
step:665/1530 train_loss:3.4897 train_time:109788ms step_avg:167.61ms
step:666/1530 train_loss:3.7754 train_time:109958ms step_avg:167.62ms
step:667/1530 train_loss:3.6015 train_time:110128ms step_avg:167.62ms
step:668/1530 train_loss:3.6409 train_time:110299ms step_avg:167.63ms
step:669/1530 train_loss:3.4863 train_time:110472ms step_avg:167.64ms
step:670/1530 train_loss:3.6005 train_time:110641ms step_avg:167.64ms
step:671/1530 train_loss:3.5612 train_time:110812ms step_avg:167.64ms
step:672/1530 train_loss:3.5632 train_time:110984ms step_avg:167.65ms
step:673/1530 train_loss:3.8437 train_time:111155ms step_avg:167.65ms
step:674/1530 train_loss:3.6176 train_time:111325ms step_avg:167.66ms
step:675/1530 train_loss:3.7065 train_time:111497ms step_avg:167.66ms
step:676/1530 train_loss:3.4889 train_time:111668ms step_avg:167.67ms
step:677/1530 train_loss:3.6008 train_time:111839ms step_avg:167.67ms
step:678/1530 train_loss:3.5520 train_time:112010ms step_avg:167.68ms
step:679/1530 train_loss:3.6737 train_time:112181ms step_avg:167.68ms
step:680/1530 train_loss:3.5840 train_time:112352ms step_avg:167.69ms
step:681/1530 train_loss:3.6099 train_time:112524ms step_avg:167.70ms
step:682/1530 train_loss:3.6603 train_time:112700ms step_avg:167.71ms
step:683/1530 train_loss:3.7279 train_time:112874ms step_avg:167.72ms
step:684/1530 train_loss:3.6423 train_time:113044ms step_avg:167.72ms
step:685/1530 train_loss:3.6783 train_time:113220ms step_avg:167.73ms
step:686/1530 train_loss:3.6351 train_time:113395ms step_avg:167.74ms
step:687/1530 train_loss:3.6576 train_time:113567ms step_avg:167.75ms
step:688/1530 train_loss:3.2019 train_time:113742ms step_avg:167.76ms
step:689/1530 train_loss:3.4017 train_time:113916ms step_avg:167.77ms
step:690/1530 train_loss:3.5401 train_time:114091ms step_avg:167.78ms
step:691/1530 train_loss:3.4085 train_time:114262ms step_avg:167.79ms
step:692/1530 train_loss:3.6260 train_time:114435ms step_avg:167.79ms
step:693/1530 train_loss:3.6414 train_time:114607ms step_avg:167.80ms
step:694/1530 train_loss:3.5492 train_time:114779ms step_avg:167.81ms
step:695/1530 train_loss:3.5270 train_time:114950ms step_avg:167.81ms
step:696/1530 train_loss:3.8451 train_time:115123ms step_avg:167.82ms
step:697/1530 train_loss:3.5837 train_time:115297ms step_avg:167.83ms
step:698/1530 train_loss:3.6428 train_time:115468ms step_avg:167.83ms
step:699/1530 train_loss:3.7631 train_time:115642ms step_avg:167.84ms
step:700/1530 train_loss:3.5715 train_time:115814ms step_avg:167.85ms
step:701/1530 train_loss:3.5354 train_time:115985ms step_avg:167.85ms
step:702/1530 train_loss:3.5084 train_time:116160ms step_avg:167.86ms
step:703/1530 train_loss:3.4988 train_time:116334ms step_avg:167.87ms
step:704/1530 train_loss:3.5653 train_time:116506ms step_avg:167.88ms
step:705/1530 train_loss:3.5559 train_time:116683ms step_avg:167.89ms
step:706/1530 train_loss:3.5769 train_time:116859ms step_avg:167.90ms
step:707/1530 train_loss:3.6425 train_time:117033ms step_avg:167.91ms
step:708/1530 train_loss:3.6038 train_time:117205ms step_avg:167.92ms
step:709/1530 train_loss:3.5761 train_time:117380ms step_avg:167.93ms
step:710/1530 train_loss:3.5340 train_time:117551ms step_avg:167.93ms
step:711/1530 train_loss:3.5874 train_time:117724ms step_avg:167.94ms
step:712/1530 train_loss:3.6408 train_time:117899ms step_avg:167.95ms
step:713/1530 train_loss:3.6517 train_time:118075ms step_avg:167.96ms
step:714/1530 train_loss:3.5549 train_time:118247ms step_avg:167.96ms
step:715/1530 train_loss:3.5671 train_time:118419ms step_avg:167.97ms
step:716/1530 train_loss:3.5803 train_time:118592ms step_avg:167.98ms
step:717/1530 train_loss:3.7058 train_time:118765ms step_avg:167.98ms
step:718/1530 train_loss:3.5928 train_time:118936ms step_avg:167.99ms
step:719/1530 train_loss:3.6732 train_time:119109ms step_avg:168.00ms
step:720/1530 train_loss:3.8441 train_time:119283ms step_avg:168.00ms
step:721/1530 train_loss:3.4616 train_time:119456ms step_avg:168.01ms
step:722/1530 train_loss:3.7340 train_time:119629ms step_avg:168.02ms
step:723/1530 train_loss:3.7680 train_time:119801ms step_avg:168.02ms
step:724/1530 train_loss:3.5655 train_time:119974ms step_avg:168.03ms
step:725/1530 train_loss:3.6457 train_time:120146ms step_avg:168.04ms
step:726/1530 train_loss:3.5257 train_time:120320ms step_avg:168.05ms
step:727/1530 train_loss:3.5719 train_time:120496ms step_avg:168.06ms
step:728/1530 train_loss:3.7242 train_time:120668ms step_avg:168.06ms
step:729/1530 train_loss:3.6604 train_time:120840ms step_avg:168.07ms
step:730/1530 train_loss:3.6581 train_time:121015ms step_avg:168.08ms
step:731/1530 train_loss:3.5522 train_time:121188ms step_avg:168.08ms
step:732/1530 train_loss:3.5870 train_time:121359ms step_avg:168.09ms
step:733/1530 train_loss:3.8280 train_time:121534ms step_avg:168.10ms
step:734/1530 train_loss:3.5560 train_time:121708ms step_avg:168.10ms
step:735/1530 train_loss:3.6121 train_time:121879ms step_avg:168.11ms
step:736/1530 train_loss:3.7321 train_time:122053ms step_avg:168.12ms
step:737/1530 train_loss:3.6712 train_time:122225ms step_avg:168.12ms
step:738/1530 train_loss:3.5970 train_time:122398ms step_avg:168.13ms
step:739/1530 train_loss:3.4971 train_time:122570ms step_avg:168.13ms
step:740/1530 train_loss:4.1069 train_time:122745ms step_avg:168.14ms
step:741/1530 train_loss:3.4806 train_time:122918ms step_avg:168.15ms
step:742/1530 train_loss:3.5542 train_time:123092ms step_avg:168.16ms
step:743/1530 train_loss:3.5759 train_time:123264ms step_avg:168.16ms
step:744/1530 train_loss:3.6456 train_time:123436ms step_avg:168.17ms
step:745/1530 train_loss:3.5875 train_time:123610ms step_avg:168.18ms
step:746/1530 train_loss:3.5976 train_time:123781ms step_avg:168.18ms
step:747/1530 train_loss:3.6401 train_time:123954ms step_avg:168.19ms
step:748/1530 train_loss:3.5606 train_time:124130ms step_avg:168.20ms
step:749/1530 train_loss:3.5609 train_time:124302ms step_avg:168.20ms
step:750/1530 train_loss:3.5875 train_time:124473ms step_avg:168.21ms
step:750/1530 val_loss:3.5604 train_time:124522ms step_avg:168.27ms
step:751/1530 train_loss:3.5656 train_time:124645ms step_avg:168.21ms
step:752/1530 train_loss:3.6094 train_time:124819ms step_avg:168.22ms
step:753/1530 train_loss:3.6173 train_time:124992ms step_avg:168.23ms
step:754/1530 train_loss:3.5907 train_time:125164ms step_avg:168.23ms
step:755/1530 train_loss:3.6819 train_time:125471ms step_avg:168.42ms
step:756/1530 train_loss:3.4540 train_time:125655ms step_avg:168.44ms
step:757/1530 train_loss:3.7200 train_time:125829ms step_avg:168.45ms
step:758/1530 train_loss:3.6504 train_time:126001ms step_avg:168.45ms
step:759/1530 train_loss:3.5849 train_time:126330ms step_avg:168.66ms
step:760/1530 train_loss:3.6972 train_time:126502ms step_avg:168.67ms
step:761/1530 train_loss:3.3951 train_time:126673ms step_avg:168.67ms
step:762/1530 train_loss:3.5459 train_time:126844ms step_avg:168.68ms
step:763/1530 train_loss:3.6632 train_time:127019ms step_avg:168.68ms
step:764/1530 train_loss:3.3152 train_time:127192ms step_avg:168.69ms
step:765/1530 train_loss:3.7223 train_time:127365ms step_avg:168.70ms
step:766/1530 train_loss:3.5608 train_time:127539ms step_avg:168.70ms
step:767/1530 train_loss:3.5583 train_time:127712ms step_avg:168.71ms
step:768/1530 train_loss:3.5679 train_time:127884ms step_avg:168.71ms
step:769/1530 train_loss:3.5830 train_time:128057ms step_avg:168.72ms
step:770/1530 train_loss:3.6384 train_time:128227ms step_avg:168.72ms
step:771/1530 train_loss:3.8833 train_time:128402ms step_avg:168.73ms
step:772/1530 train_loss:3.4436 train_time:128572ms step_avg:168.73ms
step:773/1530 train_loss:3.6287 train_time:128745ms step_avg:168.74ms
step:774/1530 train_loss:3.6401 train_time:128919ms step_avg:168.74ms
step:775/1530 train_loss:3.6016 train_time:129091ms step_avg:168.75ms
step:776/1530 train_loss:3.4004 train_time:129263ms step_avg:168.75ms
step:777/1530 train_loss:3.3844 train_time:129437ms step_avg:168.76ms
step:778/1530 train_loss:3.4889 train_time:129609ms step_avg:168.76ms
step:779/1530 train_loss:3.5799 train_time:129781ms step_avg:168.77ms
step:780/1530 train_loss:3.5849 train_time:129953ms step_avg:168.77ms
step:781/1530 train_loss:3.6756 train_time:130125ms step_avg:168.77ms
step:782/1530 train_loss:3.5840 train_time:130299ms step_avg:168.78ms
step:783/1530 train_loss:3.5572 train_time:130470ms step_avg:168.78ms
step:784/1530 train_loss:3.5985 train_time:130642ms step_avg:168.79ms
step:785/1530 train_loss:3.5533 train_time:130814ms step_avg:168.79ms
step:786/1530 train_loss:3.4350 train_time:130986ms step_avg:168.80ms
step:787/1530 train_loss:3.7458 train_time:131158ms step_avg:168.80ms
step:788/1530 train_loss:3.5003 train_time:131331ms step_avg:168.81ms
step:789/1530 train_loss:3.5418 train_time:131502ms step_avg:168.81ms
step:790/1530 train_loss:3.6241 train_time:131676ms step_avg:168.82ms
step:791/1530 train_loss:3.7725 train_time:131852ms step_avg:168.82ms
step:792/1530 train_loss:3.7587 train_time:132025ms step_avg:168.83ms
step:793/1530 train_loss:3.4417 train_time:132196ms step_avg:168.83ms
step:794/1530 train_loss:3.5900 train_time:132368ms step_avg:168.84ms
step:795/1530 train_loss:3.6675 train_time:132542ms step_avg:168.84ms
step:796/1530 train_loss:3.7451 train_time:132720ms step_avg:168.85ms
step:797/1530 train_loss:3.5242 train_time:132894ms step_avg:168.86ms
step:798/1530 train_loss:3.6422 train_time:133067ms step_avg:168.87ms
step:799/1530 train_loss:3.5278 train_time:133245ms step_avg:168.88ms
step:800/1530 train_loss:3.5242 train_time:133419ms step_avg:168.88ms
step:801/1530 train_loss:3.6194 train_time:133592ms step_avg:168.89ms
step:802/1530 train_loss:3.4955 train_time:133768ms step_avg:168.90ms
step:803/1530 train_loss:3.4796 train_time:133941ms step_avg:168.90ms
step:804/1530 train_loss:3.6158 train_time:134115ms step_avg:168.91ms
step:805/1530 train_loss:3.5155 train_time:134290ms step_avg:168.92ms
step:806/1530 train_loss:3.5598 train_time:134463ms step_avg:168.92ms
step:807/1530 train_loss:3.6361 train_time:134637ms step_avg:168.93ms
step:808/1530 train_loss:3.5430 train_time:134815ms step_avg:168.94ms
step:809/1530 train_loss:3.4855 train_time:134988ms step_avg:168.95ms
step:810/1530 train_loss:3.5574 train_time:135160ms step_avg:168.95ms
step:811/1530 train_loss:3.5808 train_time:135334ms step_avg:168.96ms
step:812/1530 train_loss:3.5909 train_time:135506ms step_avg:168.96ms
step:813/1530 train_loss:3.6202 train_time:135679ms step_avg:168.97ms
step:814/1530 train_loss:3.5619 train_time:135853ms step_avg:168.97ms
step:815/1530 train_loss:3.5613 train_time:136026ms step_avg:168.98ms
step:816/1530 train_loss:3.6824 train_time:136203ms step_avg:168.99ms
step:817/1530 train_loss:3.7624 train_time:136375ms step_avg:168.99ms
step:818/1530 train_loss:3.5149 train_time:136547ms step_avg:168.99ms
step:819/1530 train_loss:3.7061 train_time:136722ms step_avg:169.00ms
step:820/1530 train_loss:3.4880 train_time:136898ms step_avg:169.01ms
step:821/1530 train_loss:3.5652 train_time:137069ms step_avg:169.01ms
step:822/1530 train_loss:3.6940 train_time:137245ms step_avg:169.02ms
step:823/1530 train_loss:3.5702 train_time:137421ms step_avg:169.03ms
step:824/1530 train_loss:3.5064 train_time:137594ms step_avg:169.03ms
step:825/1530 train_loss:3.6103 train_time:137769ms step_avg:169.04ms
step:826/1530 train_loss:3.4735 train_time:137944ms step_avg:169.05ms
step:827/1530 train_loss:3.7260 train_time:138119ms step_avg:169.06ms
step:828/1530 train_loss:3.6135 train_time:138292ms step_avg:169.06ms
step:829/1530 train_loss:3.6223 train_time:138467ms step_avg:169.07ms
step:830/1530 train_loss:3.5287 train_time:138641ms step_avg:169.07ms
step:831/1530 train_loss:3.5966 train_time:138815ms step_avg:169.08ms
step:832/1530 train_loss:3.5094 train_time:138988ms step_avg:169.09ms
step:833/1530 train_loss:3.6559 train_time:139162ms step_avg:169.09ms
step:834/1530 train_loss:3.4749 train_time:139337ms step_avg:169.10ms
step:835/1530 train_loss:3.4562 train_time:139510ms step_avg:169.10ms
step:836/1530 train_loss:3.7132 train_time:139685ms step_avg:169.11ms
step:837/1530 train_loss:3.3899 train_time:139859ms step_avg:169.12ms
step:838/1530 train_loss:3.5875 train_time:140033ms step_avg:169.12ms
step:839/1530 train_loss:3.4171 train_time:140209ms step_avg:169.13ms
step:840/1530 train_loss:3.4694 train_time:140381ms step_avg:169.13ms
step:841/1530 train_loss:3.5669 train_time:140554ms step_avg:169.14ms
step:842/1530 train_loss:3.5811 train_time:140728ms step_avg:169.14ms
step:843/1530 train_loss:3.5560 train_time:140900ms step_avg:169.15ms
step:844/1530 train_loss:3.4232 train_time:141073ms step_avg:169.15ms
step:845/1530 train_loss:3.6520 train_time:141246ms step_avg:169.16ms
step:846/1530 train_loss:3.5103 train_time:141423ms step_avg:169.17ms
step:847/1530 train_loss:3.4932 train_time:141600ms step_avg:169.18ms
step:848/1530 train_loss:3.6387 train_time:141772ms step_avg:169.18ms
step:849/1530 train_loss:3.4828 train_time:141947ms step_avg:169.19ms
step:850/1530 train_loss:3.4428 train_time:142122ms step_avg:169.19ms
step:851/1530 train_loss:3.7334 train_time:142295ms step_avg:169.20ms
step:852/1530 train_loss:3.4390 train_time:142467ms step_avg:169.20ms
step:853/1530 train_loss:3.5635 train_time:142640ms step_avg:169.21ms
step:854/1530 train_loss:3.6505 train_time:142816ms step_avg:169.21ms
step:855/1530 train_loss:3.5116 train_time:142988ms step_avg:169.22ms
step:856/1530 train_loss:3.5453 train_time:143161ms step_avg:169.22ms
step:857/1530 train_loss:3.6008 train_time:143336ms step_avg:169.23ms
step:858/1530 train_loss:3.4641 train_time:143512ms step_avg:169.24ms
step:859/1530 train_loss:3.5491 train_time:143685ms step_avg:169.24ms
step:860/1530 train_loss:3.5773 train_time:143857ms step_avg:169.24ms
step:861/1530 train_loss:3.6336 train_time:144034ms step_avg:169.25ms
step:862/1530 train_loss:3.6065 train_time:144212ms step_avg:169.26ms
step:863/1530 train_loss:3.5683 train_time:144387ms step_avg:169.27ms
step:864/1530 train_loss:3.3799 train_time:144562ms step_avg:169.28ms
step:865/1530 train_loss:3.5937 train_time:144733ms step_avg:169.28ms
step:866/1530 train_loss:3.8738 train_time:144911ms step_avg:169.29ms
step:867/1530 train_loss:3.4576 train_time:145083ms step_avg:169.29ms
step:868/1530 train_loss:3.6435 train_time:145255ms step_avg:169.29ms
step:869/1530 train_loss:3.6132 train_time:145429ms step_avg:169.30ms
step:870/1530 train_loss:3.4461 train_time:145604ms step_avg:169.31ms
step:871/1530 train_loss:3.3898 train_time:145778ms step_avg:169.31ms
step:872/1530 train_loss:3.6411 train_time:145954ms step_avg:169.32ms
step:873/1530 train_loss:3.4573 train_time:146127ms step_avg:169.32ms
step:874/1530 train_loss:3.2209 train_time:146305ms step_avg:169.33ms
step:875/1530 train_loss:3.6297 train_time:146479ms step_avg:169.34ms
step:875/1530 val_loss:3.5154 train_time:146529ms step_avg:169.40ms
step:876/1530 train_loss:3.4390 train_time:146653ms step_avg:169.34ms
step:877/1530 train_loss:3.6156 train_time:146829ms step_avg:169.35ms
step:878/1530 train_loss:3.4597 train_time:147003ms step_avg:169.36ms
step:879/1530 train_loss:3.6422 train_time:147174ms step_avg:169.36ms
step:880/1530 train_loss:3.3002 train_time:147346ms step_avg:169.36ms
step:881/1530 train_loss:3.4713 train_time:147519ms step_avg:169.37ms
step:882/1530 train_loss:3.6918 train_time:147692ms step_avg:169.37ms
step:883/1530 train_loss:3.8322 train_time:147866ms step_avg:169.38ms
step:884/1530 train_loss:3.5634 train_time:148042ms step_avg:169.38ms
step:885/1530 train_loss:3.4923 train_time:148214ms step_avg:169.39ms
step:886/1530 train_loss:3.5667 train_time:148387ms step_avg:169.39ms
step:887/1530 train_loss:4.0850 train_time:148562ms step_avg:169.40ms
step:888/1530 train_loss:3.8324 train_time:148743ms step_avg:169.41ms
step:889/1530 train_loss:3.5141 train_time:148916ms step_avg:169.42ms
step:890/1530 train_loss:3.5260 train_time:149087ms step_avg:169.42ms
step:891/1530 train_loss:3.3565 train_time:149263ms step_avg:169.42ms
step:892/1530 train_loss:3.7136 train_time:149436ms step_avg:169.43ms
step:893/1530 train_loss:3.4188 train_time:149607ms step_avg:169.43ms
step:894/1530 train_loss:3.6425 train_time:149783ms step_avg:169.44ms
step:895/1530 train_loss:3.6755 train_time:149958ms step_avg:169.44ms
step:896/1530 train_loss:3.4912 train_time:150132ms step_avg:169.45ms
step:897/1530 train_loss:3.5386 train_time:150307ms step_avg:169.46ms
step:898/1530 train_loss:3.5878 train_time:150483ms step_avg:169.46ms
step:899/1530 train_loss:3.4745 train_time:150655ms step_avg:169.47ms
step:900/1530 train_loss:3.4220 train_time:150826ms step_avg:169.47ms
step:901/1530 train_loss:3.6146 train_time:151000ms step_avg:169.47ms
step:902/1530 train_loss:3.6299 train_time:151171ms step_avg:169.47ms
step:903/1530 train_loss:3.5350 train_time:151349ms step_avg:169.48ms
step:904/1530 train_loss:3.4952 train_time:151524ms step_avg:169.49ms
step:905/1530 train_loss:3.4990 train_time:151694ms step_avg:169.49ms
step:906/1530 train_loss:3.6980 train_time:151868ms step_avg:169.50ms
step:907/1530 train_loss:3.5060 train_time:152044ms step_avg:169.50ms
step:908/1530 train_loss:3.5621 train_time:152216ms step_avg:169.51ms
step:909/1530 train_loss:3.4548 train_time:152392ms step_avg:169.51ms
step:910/1530 train_loss:3.5223 train_time:152569ms step_avg:169.52ms
step:911/1530 train_loss:3.6380 train_time:152747ms step_avg:169.53ms
step:912/1530 train_loss:3.5962 train_time:152926ms step_avg:169.54ms
step:913/1530 train_loss:3.4588 train_time:153104ms step_avg:169.55ms
step:914/1530 train_loss:3.7419 train_time:153283ms step_avg:169.56ms
step:915/1530 train_loss:3.5351 train_time:153462ms step_avg:169.57ms
step:916/1530 train_loss:3.6134 train_time:153639ms step_avg:169.58ms
step:917/1530 train_loss:3.5942 train_time:153812ms step_avg:169.58ms
step:918/1530 train_loss:4.8055 train_time:153991ms step_avg:169.59ms
step:919/1530 train_loss:3.4915 train_time:154169ms step_avg:169.60ms
step:920/1530 train_loss:3.5806 train_time:154343ms step_avg:169.61ms
step:921/1530 train_loss:3.5504 train_time:154519ms step_avg:169.61ms
step:922/1530 train_loss:3.5795 train_time:154696ms step_avg:169.62ms
step:923/1530 train_loss:3.6045 train_time:154870ms step_avg:169.63ms
step:924/1530 train_loss:3.6796 train_time:155048ms step_avg:169.64ms
step:925/1530 train_loss:3.6441 train_time:155223ms step_avg:169.64ms
step:926/1530 train_loss:3.5553 train_time:155397ms step_avg:169.65ms
step:927/1530 train_loss:3.5506 train_time:155571ms step_avg:169.65ms
step:928/1530 train_loss:3.7749 train_time:155749ms step_avg:169.66ms
step:929/1530 train_loss:3.6052 train_time:155925ms step_avg:169.67ms
step:930/1530 train_loss:3.3949 train_time:156102ms step_avg:169.68ms
step:931/1530 train_loss:3.4961 train_time:156275ms step_avg:169.68ms
step:932/1530 train_loss:3.6443 train_time:156454ms step_avg:169.69ms
step:933/1530 train_loss:3.3548 train_time:156630ms step_avg:169.70ms
step:934/1530 train_loss:3.5803 train_time:156807ms step_avg:169.70ms
step:935/1530 train_loss:3.4347 train_time:156985ms step_avg:169.71ms
step:936/1530 train_loss:3.5159 train_time:157163ms step_avg:169.72ms
step:937/1530 train_loss:3.6153 train_time:157341ms step_avg:169.73ms
step:938/1530 train_loss:3.5405 train_time:157515ms step_avg:169.74ms
step:939/1530 train_loss:3.6753 train_time:157695ms step_avg:169.75ms
step:940/1530 train_loss:3.4757 train_time:157870ms step_avg:169.75ms
step:941/1530 train_loss:3.5498 train_time:158046ms step_avg:169.76ms
step:942/1530 train_loss:3.3534 train_time:158222ms step_avg:169.77ms
step:943/1530 train_loss:3.7087 train_time:158401ms step_avg:169.78ms
step:944/1530 train_loss:3.4003 train_time:158715ms step_avg:169.93ms
step:945/1530 train_loss:3.4185 train_time:158899ms step_avg:169.95ms
step:946/1530 train_loss:5.0608 train_time:159077ms step_avg:169.95ms
step:947/1530 train_loss:3.5932 train_time:159252ms step_avg:169.96ms
step:948/1530 train_loss:3.4807 train_time:159427ms step_avg:169.97ms
step:949/1530 train_loss:3.3699 train_time:159759ms step_avg:170.14ms
step:950/1530 train_loss:3.4364 train_time:159936ms step_avg:170.14ms
step:951/1530 train_loss:3.4074 train_time:160113ms step_avg:170.15ms
step:952/1530 train_loss:3.4708 train_time:160290ms step_avg:170.16ms
step:953/1530 train_loss:3.5648 train_time:160469ms step_avg:170.17ms
step:954/1530 train_loss:3.4461 train_time:160648ms step_avg:170.18ms
step:955/1530 train_loss:3.4798 train_time:160823ms step_avg:170.18ms
step:956/1530 train_loss:3.4417 train_time:160997ms step_avg:170.19ms
step:957/1530 train_loss:3.4903 train_time:161174ms step_avg:170.19ms
step:958/1530 train_loss:3.5057 train_time:161355ms step_avg:170.21ms
step:959/1530 train_loss:3.5103 train_time:161530ms step_avg:170.21ms
step:960/1530 train_loss:3.4050 train_time:161707ms step_avg:170.22ms
step:961/1530 train_loss:3.6348 train_time:161882ms step_avg:170.22ms
step:962/1530 train_loss:3.5910 train_time:162057ms step_avg:170.23ms
step:963/1530 train_loss:3.6884 train_time:162234ms step_avg:170.23ms
step:964/1530 train_loss:3.4291 train_time:162412ms step_avg:170.24ms
step:965/1530 train_loss:3.4795 train_time:162585ms step_avg:170.25ms
step:966/1530 train_loss:3.7075 train_time:162761ms step_avg:170.25ms
step:967/1530 train_loss:3.5213 train_time:162937ms step_avg:170.26ms
step:968/1530 train_loss:3.5116 train_time:163113ms step_avg:170.26ms
step:969/1530 train_loss:3.5830 train_time:163287ms step_avg:170.27ms
step:970/1530 train_loss:3.3728 train_time:163460ms step_avg:170.27ms
step:971/1530 train_loss:3.5305 train_time:163637ms step_avg:170.28ms
step:972/1530 train_loss:3.4828 train_time:163810ms step_avg:170.28ms
step:973/1530 train_loss:3.5367 train_time:163984ms step_avg:170.28ms
step:974/1530 train_loss:3.5904 train_time:164161ms step_avg:170.29ms
step:975/1530 train_loss:3.4614 train_time:164338ms step_avg:170.30ms
step:976/1530 train_loss:3.6648 train_time:164512ms step_avg:170.30ms
step:977/1530 train_loss:3.5681 train_time:164685ms step_avg:170.30ms
step:978/1530 train_loss:3.3505 train_time:164860ms step_avg:170.31ms
step:979/1530 train_loss:3.6294 train_time:165036ms step_avg:170.32ms
step:980/1530 train_loss:3.4204 train_time:165213ms step_avg:170.32ms
step:981/1530 train_loss:3.5724 train_time:165390ms step_avg:170.33ms
step:982/1530 train_loss:3.5407 train_time:165564ms step_avg:170.33ms
step:983/1530 train_loss:3.5075 train_time:165743ms step_avg:170.34ms
step:984/1530 train_loss:3.4980 train_time:165916ms step_avg:170.35ms
step:985/1530 train_loss:3.5743 train_time:166091ms step_avg:170.35ms
step:986/1530 train_loss:3.4132 train_time:166266ms step_avg:170.35ms
step:987/1530 train_loss:3.4756 train_time:166440ms step_avg:170.36ms
step:988/1530 train_loss:3.4948 train_time:166613ms step_avg:170.36ms
step:989/1530 train_loss:3.4159 train_time:166785ms step_avg:170.36ms
step:990/1530 train_loss:3.6611 train_time:166962ms step_avg:170.37ms
step:991/1530 train_loss:3.4670 train_time:167137ms step_avg:170.37ms
step:992/1530 train_loss:3.4421 train_time:167316ms step_avg:170.38ms
step:993/1530 train_loss:3.4957 train_time:167495ms step_avg:170.39ms
step:994/1530 train_loss:3.5974 train_time:167669ms step_avg:170.40ms
step:995/1530 train_loss:3.5255 train_time:167842ms step_avg:170.40ms
step:996/1530 train_loss:3.4551 train_time:168015ms step_avg:170.40ms
step:997/1530 train_loss:3.7523 train_time:168186ms step_avg:170.40ms
step:998/1530 train_loss:3.4378 train_time:168360ms step_avg:170.40ms
step:999/1530 train_loss:3.5840 train_time:168535ms step_avg:170.41ms
step:1000/1530 train_loss:3.4352 train_time:168713ms step_avg:170.42ms
step:1000/1530 val_loss:3.4634 train_time:168766ms step_avg:170.47ms
step:1001/1530 train_loss:3.4965 train_time:168891ms step_avg:170.42ms
step:1002/1530 train_loss:3.3705 train_time:169064ms step_avg:170.43ms
step:1003/1530 train_loss:3.5554 train_time:169241ms step_avg:170.43ms
step:1004/1530 train_loss:3.6044 train_time:169416ms step_avg:170.44ms
step:1005/1530 train_loss:3.3883 train_time:169591ms step_avg:170.44ms
step:1006/1530 train_loss:3.4638 train_time:169769ms step_avg:170.45ms
step:1007/1530 train_loss:3.4366 train_time:169945ms step_avg:170.46ms
step:1008/1530 train_loss:3.5593 train_time:170119ms step_avg:170.46ms
step:1009/1530 train_loss:3.6579 train_time:170296ms step_avg:170.47ms
step:1010/1530 train_loss:3.5628 train_time:170469ms step_avg:170.47ms
step:1011/1530 train_loss:3.5343 train_time:170643ms step_avg:170.47ms
step:1012/1530 train_loss:3.3893 train_time:170816ms step_avg:170.47ms
step:1013/1530 train_loss:3.5364 train_time:170991ms step_avg:170.48ms
step:1014/1530 train_loss:3.6195 train_time:171169ms step_avg:170.49ms
step:1015/1530 train_loss:3.3265 train_time:171348ms step_avg:170.50ms
step:1016/1530 train_loss:3.4090 train_time:171522ms step_avg:170.50ms
step:1017/1530 train_loss:3.3932 train_time:171698ms step_avg:170.50ms
step:1018/1530 train_loss:3.3955 train_time:171874ms step_avg:170.51ms
step:1019/1530 train_loss:3.5191 train_time:172050ms step_avg:170.51ms
step:1020/1530 train_loss:3.3785 train_time:172224ms step_avg:170.52ms
step:1021/1530 train_loss:3.3499 train_time:172398ms step_avg:170.52ms
step:1022/1530 train_loss:3.4790 train_time:172575ms step_avg:170.53ms
step:1023/1530 train_loss:3.5042 train_time:172753ms step_avg:170.54ms
step:1024/1530 train_loss:3.4781 train_time:172931ms step_avg:170.54ms
step:1025/1530 train_loss:3.4741 train_time:173108ms step_avg:170.55ms
step:1026/1530 train_loss:3.6176 train_time:173285ms step_avg:170.56ms
step:1027/1530 train_loss:3.3191 train_time:173460ms step_avg:170.56ms
step:1028/1530 train_loss:3.3966 train_time:173642ms step_avg:170.57ms
step:1029/1530 train_loss:3.3145 train_time:173822ms step_avg:170.58ms
step:1030/1530 train_loss:3.5336 train_time:173998ms step_avg:170.59ms
step:1031/1530 train_loss:3.5122 train_time:174174ms step_avg:170.59ms
step:1032/1530 train_loss:3.6911 train_time:174357ms step_avg:170.60ms
step:1033/1530 train_loss:3.4906 train_time:174532ms step_avg:170.61ms
step:1034/1530 train_loss:3.3927 train_time:174708ms step_avg:170.61ms
step:1035/1530 train_loss:3.4442 train_time:174887ms step_avg:170.62ms
step:1036/1530 train_loss:3.4789 train_time:175064ms step_avg:170.63ms
step:1037/1530 train_loss:3.7854 train_time:175242ms step_avg:170.64ms
step:1038/1530 train_loss:3.6170 train_time:175420ms step_avg:170.64ms
step:1039/1530 train_loss:3.5085 train_time:175601ms step_avg:170.65ms
step:1040/1530 train_loss:3.4098 train_time:175776ms step_avg:170.66ms
step:1041/1530 train_loss:3.4813 train_time:175955ms step_avg:170.66ms
step:1042/1530 train_loss:3.5183 train_time:176129ms step_avg:170.67ms
step:1043/1530 train_loss:3.4428 train_time:176304ms step_avg:170.67ms
step:1044/1530 train_loss:3.4554 train_time:176481ms step_avg:170.68ms
step:1045/1530 train_loss:3.5140 train_time:176661ms step_avg:170.69ms
step:1046/1530 train_loss:3.4239 train_time:176838ms step_avg:170.69ms
step:1047/1530 train_loss:3.6320 train_time:177014ms step_avg:170.70ms
step:1048/1530 train_loss:3.4877 train_time:177191ms step_avg:170.70ms
step:1049/1530 train_loss:3.3967 train_time:177366ms step_avg:170.71ms
step:1050/1530 train_loss:3.3851 train_time:177543ms step_avg:170.71ms
step:1051/1530 train_loss:3.4889 train_time:177719ms step_avg:170.72ms
step:1052/1530 train_loss:3.3604 train_time:177896ms step_avg:170.73ms
step:1053/1530 train_loss:3.6884 train_time:178075ms step_avg:170.73ms
step:1054/1530 train_loss:3.5361 train_time:178254ms step_avg:170.74ms
step:1055/1530 train_loss:3.3827 train_time:178430ms step_avg:170.75ms
step:1056/1530 train_loss:3.4981 train_time:178605ms step_avg:170.75ms
step:1057/1530 train_loss:3.5784 train_time:178780ms step_avg:170.75ms
step:1058/1530 train_loss:3.3007 train_time:178959ms step_avg:170.76ms
step:1059/1530 train_loss:3.3629 train_time:179141ms step_avg:170.77ms
step:1060/1530 train_loss:3.4381 train_time:179317ms step_avg:170.78ms
step:1061/1530 train_loss:3.4149 train_time:179491ms step_avg:170.78ms
step:1062/1530 train_loss:3.3822 train_time:179668ms step_avg:170.79ms
step:1063/1530 train_loss:3.4541 train_time:179844ms step_avg:170.79ms
step:1064/1530 train_loss:3.3800 train_time:180017ms step_avg:170.79ms
step:1065/1530 train_loss:3.3546 train_time:180194ms step_avg:170.80ms
step:1066/1530 train_loss:3.4117 train_time:180372ms step_avg:170.81ms
step:1067/1530 train_loss:3.2762 train_time:180550ms step_avg:170.81ms
step:1068/1530 train_loss:3.4323 train_time:180726ms step_avg:170.82ms
step:1069/1530 train_loss:3.2926 train_time:180906ms step_avg:170.83ms
step:1070/1530 train_loss:3.5684 train_time:181082ms step_avg:170.83ms
step:1071/1530 train_loss:3.5099 train_time:181260ms step_avg:170.84ms
step:1072/1530 train_loss:3.4327 train_time:181436ms step_avg:170.84ms
step:1073/1530 train_loss:3.5236 train_time:181609ms step_avg:170.85ms
step:1074/1530 train_loss:3.4289 train_time:181786ms step_avg:170.85ms
step:1075/1530 train_loss:3.3998 train_time:181964ms step_avg:170.86ms
step:1076/1530 train_loss:3.7933 train_time:182140ms step_avg:170.86ms
step:1077/1530 train_loss:3.4243 train_time:182316ms step_avg:170.87ms
step:1078/1530 train_loss:3.0846 train_time:182499ms step_avg:170.88ms
step:1079/1530 train_loss:3.5337 train_time:182674ms step_avg:170.88ms
step:1080/1530 train_loss:3.4255 train_time:182854ms step_avg:170.89ms
step:1081/1530 train_loss:3.4922 train_time:183029ms step_avg:170.90ms
step:1082/1530 train_loss:3.5858 train_time:183204ms step_avg:170.90ms
step:1083/1530 train_loss:3.4938 train_time:183378ms step_avg:170.90ms
step:1084/1530 train_loss:3.4591 train_time:183554ms step_avg:170.91ms
step:1085/1530 train_loss:3.4288 train_time:183731ms step_avg:170.91ms
step:1086/1530 train_loss:3.6275 train_time:183905ms step_avg:170.92ms
step:1087/1530 train_loss:3.5010 train_time:184081ms step_avg:170.92ms
step:1088/1530 train_loss:3.3679 train_time:184259ms step_avg:170.93ms
step:1089/1530 train_loss:3.3692 train_time:184439ms step_avg:170.94ms
step:1090/1530 train_loss:3.4797 train_time:184618ms step_avg:170.94ms
step:1091/1530 train_loss:3.2791 train_time:184795ms step_avg:170.95ms
step:1092/1530 train_loss:3.4789 train_time:184973ms step_avg:170.95ms
step:1093/1530 train_loss:3.6004 train_time:185153ms step_avg:170.96ms
step:1094/1530 train_loss:3.4456 train_time:185329ms step_avg:170.97ms
step:1095/1530 train_loss:3.4130 train_time:185502ms step_avg:170.97ms
step:1096/1530 train_loss:3.4227 train_time:185683ms step_avg:170.98ms
step:1097/1530 train_loss:3.4847 train_time:185859ms step_avg:170.98ms
step:1098/1530 train_loss:3.5602 train_time:186038ms step_avg:170.99ms
step:1099/1530 train_loss:3.5273 train_time:186215ms step_avg:171.00ms
step:1100/1530 train_loss:3.4212 train_time:186394ms step_avg:171.00ms
step:1101/1530 train_loss:3.2880 train_time:186571ms step_avg:171.01ms
step:1102/1530 train_loss:3.3089 train_time:186751ms step_avg:171.02ms
step:1103/1530 train_loss:3.4421 train_time:186933ms step_avg:171.03ms
step:1104/1530 train_loss:3.3208 train_time:187109ms step_avg:171.03ms
step:1105/1530 train_loss:4.0565 train_time:187289ms step_avg:171.04ms
step:1106/1530 train_loss:3.2214 train_time:187463ms step_avg:171.04ms
step:1107/1530 train_loss:3.5652 train_time:187638ms step_avg:171.05ms
step:1108/1530 train_loss:3.3400 train_time:187812ms step_avg:171.05ms
step:1109/1530 train_loss:3.4970 train_time:187988ms step_avg:171.05ms
step:1110/1530 train_loss:3.4201 train_time:188161ms step_avg:171.06ms
step:1111/1530 train_loss:3.4804 train_time:188336ms step_avg:171.06ms
step:1112/1530 train_loss:3.5570 train_time:188515ms step_avg:171.07ms
step:1113/1530 train_loss:3.4310 train_time:188698ms step_avg:171.08ms
step:1114/1530 train_loss:3.3678 train_time:188879ms step_avg:171.09ms
step:1115/1530 train_loss:3.2331 train_time:189059ms step_avg:171.09ms
step:1116/1530 train_loss:3.4210 train_time:189233ms step_avg:171.10ms
step:1117/1530 train_loss:3.5891 train_time:189412ms step_avg:171.10ms
step:1118/1530 train_loss:3.6196 train_time:189588ms step_avg:171.11ms
step:1119/1530 train_loss:3.4757 train_time:189761ms step_avg:171.11ms
step:1120/1530 train_loss:3.4950 train_time:189938ms step_avg:171.12ms
step:1121/1530 train_loss:3.3877 train_time:190115ms step_avg:171.12ms
step:1122/1530 train_loss:3.4546 train_time:190291ms step_avg:171.12ms
step:1123/1530 train_loss:3.5789 train_time:190466ms step_avg:171.13ms
step:1124/1530 train_loss:3.3398 train_time:190643ms step_avg:171.13ms
step:1125/1530 train_loss:3.2274 train_time:190818ms step_avg:171.14ms
step:1125/1530 val_loss:3.4070 train_time:190869ms step_avg:171.18ms
step:1126/1530 train_loss:3.4736 train_time:190997ms step_avg:171.14ms
step:1127/1530 train_loss:3.6694 train_time:191177ms step_avg:171.15ms
step:1128/1530 train_loss:3.2274 train_time:191353ms step_avg:171.16ms
step:1129/1530 train_loss:3.5558 train_time:191533ms step_avg:171.16ms
step:1130/1530 train_loss:3.3758 train_time:191711ms step_avg:171.17ms
step:1131/1530 train_loss:3.3951 train_time:191894ms step_avg:171.18ms
step:1132/1530 train_loss:3.3647 train_time:192067ms step_avg:171.18ms
step:1133/1530 train_loss:3.4921 train_time:192379ms step_avg:171.31ms
step:1134/1530 train_loss:3.4392 train_time:192562ms step_avg:171.32ms
step:1135/1530 train_loss:3.5140 train_time:192740ms step_avg:171.32ms
step:1136/1530 train_loss:3.5606 train_time:192918ms step_avg:171.33ms
step:1137/1530 train_loss:3.4569 train_time:193094ms step_avg:171.33ms
step:1138/1530 train_loss:3.3476 train_time:193274ms step_avg:171.34ms
step:1139/1530 train_loss:3.6524 train_time:193606ms step_avg:171.48ms
step:1140/1530 train_loss:3.4513 train_time:193783ms step_avg:171.49ms
step:1141/1530 train_loss:3.5970 train_time:193966ms step_avg:171.50ms
step:1142/1530 train_loss:3.4408 train_time:194146ms step_avg:171.51ms
step:1143/1530 train_loss:3.3583 train_time:194325ms step_avg:171.51ms
step:1144/1530 train_loss:3.4396 train_time:194502ms step_avg:171.52ms
step:1145/1530 train_loss:3.5864 train_time:194677ms step_avg:171.52ms
step:1146/1530 train_loss:3.5596 train_time:194858ms step_avg:171.53ms
step:1147/1530 train_loss:3.4844 train_time:195037ms step_avg:171.54ms
step:1148/1530 train_loss:3.4981 train_time:195214ms step_avg:171.54ms
step:1149/1530 train_loss:3.3195 train_time:195396ms step_avg:171.55ms
step:1150/1530 train_loss:3.3712 train_time:195571ms step_avg:171.55ms
step:1151/1530 train_loss:3.3155 train_time:195749ms step_avg:171.56ms
step:1152/1530 train_loss:3.3911 train_time:195930ms step_avg:171.57ms
step:1153/1530 train_loss:3.4305 train_time:196109ms step_avg:171.57ms
step:1154/1530 train_loss:3.5164 train_time:196284ms step_avg:171.58ms
step:1155/1530 train_loss:3.3183 train_time:196466ms step_avg:171.59ms
step:1156/1530 train_loss:3.5399 train_time:196650ms step_avg:171.60ms
step:1157/1530 train_loss:3.4943 train_time:196829ms step_avg:171.60ms
step:1158/1530 train_loss:3.2453 train_time:197005ms step_avg:171.61ms
step:1159/1530 train_loss:3.3473 train_time:197183ms step_avg:171.61ms
step:1160/1530 train_loss:3.3350 train_time:197359ms step_avg:171.62ms
step:1161/1530 train_loss:3.0779 train_time:197538ms step_avg:171.62ms
step:1162/1530 train_loss:3.4244 train_time:197716ms step_avg:171.63ms
step:1163/1530 train_loss:3.3873 train_time:197895ms step_avg:171.63ms
step:1164/1530 train_loss:3.2921 train_time:198070ms step_avg:171.64ms
step:1165/1530 train_loss:3.2443 train_time:198245ms step_avg:171.64ms
step:1166/1530 train_loss:3.3874 train_time:198424ms step_avg:171.65ms
step:1167/1530 train_loss:3.4063 train_time:198599ms step_avg:171.65ms
step:1168/1530 train_loss:3.7197 train_time:198774ms step_avg:171.65ms
step:1169/1530 train_loss:3.3725 train_time:198950ms step_avg:171.66ms
step:1170/1530 train_loss:3.3846 train_time:199126ms step_avg:171.66ms
step:1171/1530 train_loss:3.3177 train_time:199302ms step_avg:171.66ms
step:1172/1530 train_loss:3.4200 train_time:199477ms step_avg:171.67ms
step:1173/1530 train_loss:3.5339 train_time:199658ms step_avg:171.68ms
step:1174/1530 train_loss:3.3824 train_time:199843ms step_avg:171.69ms
step:1175/1530 train_loss:3.3597 train_time:200022ms step_avg:171.69ms
step:1176/1530 train_loss:3.4204 train_time:200203ms step_avg:171.70ms
step:1177/1530 train_loss:3.4478 train_time:200384ms step_avg:171.71ms
step:1178/1530 train_loss:3.4907 train_time:200561ms step_avg:171.71ms
step:1179/1530 train_loss:3.3975 train_time:200737ms step_avg:171.72ms
step:1180/1530 train_loss:3.3449 train_time:200925ms step_avg:171.73ms
step:1181/1530 train_loss:3.3358 train_time:201102ms step_avg:171.74ms
step:1182/1530 train_loss:3.3731 train_time:201280ms step_avg:171.74ms
step:1183/1530 train_loss:3.3338 train_time:201458ms step_avg:171.75ms
step:1184/1530 train_loss:3.5119 train_time:201634ms step_avg:171.75ms
step:1185/1530 train_loss:3.5392 train_time:201815ms step_avg:171.76ms
step:1186/1530 train_loss:3.3636 train_time:201995ms step_avg:171.76ms
step:1187/1530 train_loss:3.4178 train_time:202183ms step_avg:171.78ms
step:1188/1530 train_loss:3.4425 train_time:202359ms step_avg:171.78ms
step:1189/1530 train_loss:3.2745 train_time:202538ms step_avg:171.79ms
step:1190/1530 train_loss:3.4404 train_time:202714ms step_avg:171.79ms
step:1191/1530 train_loss:3.5817 train_time:202895ms step_avg:171.80ms
step:1192/1530 train_loss:3.3903 train_time:203069ms step_avg:171.80ms
step:1193/1530 train_loss:3.2708 train_time:203245ms step_avg:171.80ms
step:1194/1530 train_loss:3.5521 train_time:203421ms step_avg:171.81ms
step:1195/1530 train_loss:3.3700 train_time:203603ms step_avg:171.82ms
step:1196/1530 train_loss:3.3838 train_time:203789ms step_avg:171.83ms
step:1197/1530 train_loss:3.2940 train_time:203968ms step_avg:171.84ms
step:1198/1530 train_loss:3.2996 train_time:204155ms step_avg:171.85ms
step:1199/1530 train_loss:3.3422 train_time:204334ms step_avg:171.85ms
step:1200/1530 train_loss:3.4471 train_time:204509ms step_avg:171.86ms
step:1201/1530 train_loss:3.4813 train_time:204688ms step_avg:171.86ms
step:1202/1530 train_loss:3.6632 train_time:204876ms step_avg:171.88ms
step:1203/1530 train_loss:3.4065 train_time:205056ms step_avg:171.88ms
step:1204/1530 train_loss:3.3041 train_time:205236ms step_avg:171.89ms
step:1205/1530 train_loss:3.4377 train_time:205411ms step_avg:171.89ms
step:1206/1530 train_loss:3.4764 train_time:205587ms step_avg:171.90ms
step:1207/1530 train_loss:3.5119 train_time:205766ms step_avg:171.90ms
step:1208/1530 train_loss:3.3941 train_time:205941ms step_avg:171.90ms
step:1209/1530 train_loss:3.2483 train_time:206120ms step_avg:171.91ms
step:1210/1530 train_loss:3.3007 train_time:206299ms step_avg:171.92ms
step:1211/1530 train_loss:3.3937 train_time:206474ms step_avg:171.92ms
step:1212/1530 train_loss:3.3961 train_time:206651ms step_avg:171.92ms
step:1213/1530 train_loss:3.4081 train_time:206830ms step_avg:171.93ms
step:1214/1530 train_loss:3.2460 train_time:207011ms step_avg:171.94ms
step:1215/1530 train_loss:3.3931 train_time:207187ms step_avg:171.94ms
step:1216/1530 train_loss:3.3305 train_time:207364ms step_avg:171.94ms
step:1217/1530 train_loss:3.3250 train_time:207541ms step_avg:171.95ms
step:1218/1530 train_loss:3.4039 train_time:207721ms step_avg:171.95ms
step:1219/1530 train_loss:3.2535 train_time:207903ms step_avg:171.96ms
step:1220/1530 train_loss:3.4707 train_time:208079ms step_avg:171.97ms
step:1221/1530 train_loss:3.5016 train_time:208256ms step_avg:171.97ms
step:1222/1530 train_loss:3.4281 train_time:208428ms step_avg:171.97ms
step:1223/1530 train_loss:3.3008 train_time:208607ms step_avg:171.98ms
step:1224/1530 train_loss:3.2506 train_time:208788ms step_avg:171.98ms
step:1225/1530 train_loss:3.3629 train_time:208966ms step_avg:171.99ms
step:1226/1530 train_loss:3.3300 train_time:209145ms step_avg:171.99ms
step:1227/1530 train_loss:3.2756 train_time:209325ms step_avg:172.00ms
step:1228/1530 train_loss:3.4442 train_time:209501ms step_avg:172.00ms
step:1229/1530 train_loss:3.3722 train_time:209681ms step_avg:172.01ms
step:1230/1530 train_loss:3.3963 train_time:209863ms step_avg:172.02ms
step:1231/1530 train_loss:3.5747 train_time:210043ms step_avg:172.03ms
step:1232/1530 train_loss:3.4977 train_time:210222ms step_avg:172.03ms
step:1233/1530 train_loss:3.4278 train_time:210397ms step_avg:172.03ms
step:1234/1530 train_loss:3.5878 train_time:210575ms step_avg:172.04ms
step:1235/1530 train_loss:3.3198 train_time:210758ms step_avg:172.05ms
step:1236/1530 train_loss:3.2876 train_time:210933ms step_avg:172.05ms
step:1237/1530 train_loss:3.2690 train_time:211110ms step_avg:172.05ms
step:1238/1530 train_loss:3.2778 train_time:211295ms step_avg:172.06ms
step:1239/1530 train_loss:3.3335 train_time:211473ms step_avg:172.07ms
step:1240/1530 train_loss:3.3808 train_time:211649ms step_avg:172.07ms
step:1241/1530 train_loss:3.4247 train_time:211828ms step_avg:172.08ms
step:1242/1530 train_loss:3.2988 train_time:212005ms step_avg:172.08ms
step:1243/1530 train_loss:3.4022 train_time:212184ms step_avg:172.09ms
step:1244/1530 train_loss:3.4038 train_time:212358ms step_avg:172.09ms
step:1245/1530 train_loss:3.4073 train_time:212535ms step_avg:172.09ms
step:1246/1530 train_loss:3.2447 train_time:212713ms step_avg:172.10ms
step:1247/1530 train_loss:3.3708 train_time:212888ms step_avg:172.10ms
step:1248/1530 train_loss:3.4229 train_time:213065ms step_avg:172.10ms
step:1249/1530 train_loss:3.4262 train_time:213243ms step_avg:172.11ms
step:1250/1530 train_loss:3.3015 train_time:213421ms step_avg:172.11ms
step:1250/1530 val_loss:3.3537 train_time:213474ms step_avg:172.16ms
step:1251/1530 train_loss:3.4864 train_time:213605ms step_avg:172.12ms
step:1252/1530 train_loss:3.3583 train_time:213781ms step_avg:172.13ms
step:1253/1530 train_loss:3.3031 train_time:213958ms step_avg:172.13ms
step:1254/1530 train_loss:3.4075 train_time:214140ms step_avg:172.14ms
step:1255/1530 train_loss:3.5158 train_time:214329ms step_avg:172.15ms
step:1256/1530 train_loss:3.3053 train_time:214512ms step_avg:172.16ms
step:1257/1530 train_loss:3.3748 train_time:214691ms step_avg:172.17ms
step:1258/1530 train_loss:3.3617 train_time:214874ms step_avg:172.17ms
step:1259/1530 train_loss:3.3286 train_time:215053ms step_avg:172.18ms
step:1260/1530 train_loss:3.2107 train_time:215229ms step_avg:172.18ms
step:1261/1530 train_loss:3.3084 train_time:215409ms step_avg:172.19ms
step:1262/1530 train_loss:3.3212 train_time:215590ms step_avg:172.20ms
step:1263/1530 train_loss:3.2365 train_time:215772ms step_avg:172.20ms
step:1264/1530 train_loss:3.4420 train_time:215946ms step_avg:172.21ms
step:1265/1530 train_loss:3.4200 train_time:216122ms step_avg:172.21ms
step:1266/1530 train_loss:3.4387 train_time:216301ms step_avg:172.21ms
step:1267/1530 train_loss:3.3717 train_time:216483ms step_avg:172.22ms
step:1268/1530 train_loss:3.4101 train_time:216663ms step_avg:172.23ms
step:1269/1530 train_loss:3.2550 train_time:216848ms step_avg:172.24ms
step:1270/1530 train_loss:3.1042 train_time:217027ms step_avg:172.24ms
step:1271/1530 train_loss:3.4024 train_time:217205ms step_avg:172.25ms
step:1272/1530 train_loss:3.3507 train_time:217384ms step_avg:172.25ms
step:1273/1530 train_loss:3.3765 train_time:217565ms step_avg:172.26ms
step:1274/1530 train_loss:3.3629 train_time:217746ms step_avg:172.27ms
step:1275/1530 train_loss:3.4309 train_time:217921ms step_avg:172.27ms
step:1276/1530 train_loss:3.4678 train_time:218097ms step_avg:172.27ms
step:1277/1530 train_loss:3.4111 train_time:218276ms step_avg:172.28ms
step:1278/1530 train_loss:3.4084 train_time:218453ms step_avg:172.28ms
step:1279/1530 train_loss:3.2680 train_time:218634ms step_avg:172.29ms
step:1280/1530 train_loss:3.3639 train_time:218816ms step_avg:172.30ms
step:1281/1530 train_loss:3.4182 train_time:218993ms step_avg:172.30ms
step:1282/1530 train_loss:3.4638 train_time:219168ms step_avg:172.30ms
step:1283/1530 train_loss:3.3364 train_time:219347ms step_avg:172.31ms
step:1284/1530 train_loss:3.3718 train_time:219526ms step_avg:172.31ms
step:1285/1530 train_loss:3.3612 train_time:219704ms step_avg:172.32ms
step:1286/1530 train_loss:3.3345 train_time:219881ms step_avg:172.32ms
step:1287/1530 train_loss:3.4890 train_time:220060ms step_avg:172.33ms
step:1288/1530 train_loss:3.2965 train_time:220241ms step_avg:172.33ms
step:1289/1530 train_loss:3.3784 train_time:220429ms step_avg:172.34ms
step:1290/1530 train_loss:3.4616 train_time:220615ms step_avg:172.36ms
step:1291/1530 train_loss:3.3821 train_time:220794ms step_avg:172.36ms
step:1292/1530 train_loss:3.4761 train_time:220976ms step_avg:172.37ms
step:1293/1530 train_loss:3.5128 train_time:221156ms step_avg:172.37ms
step:1294/1530 train_loss:3.4535 train_time:221337ms step_avg:172.38ms
step:1295/1530 train_loss:3.2835 train_time:221517ms step_avg:172.39ms
step:1296/1530 train_loss:3.3703 train_time:221698ms step_avg:172.39ms
step:1297/1530 train_loss:3.2744 train_time:221880ms step_avg:172.40ms
step:1298/1530 train_loss:3.2710 train_time:222061ms step_avg:172.41ms
step:1299/1530 train_loss:3.3975 train_time:222239ms step_avg:172.41ms
step:1300/1530 train_loss:3.3999 train_time:222416ms step_avg:172.42ms
step:1301/1530 train_loss:3.4001 train_time:222593ms step_avg:172.42ms
step:1302/1530 train_loss:3.5764 train_time:222775ms step_avg:172.43ms
step:1303/1530 train_loss:3.3019 train_time:222958ms step_avg:172.43ms
step:1304/1530 train_loss:3.5125 train_time:223139ms step_avg:172.44ms
step:1305/1530 train_loss:3.2575 train_time:223315ms step_avg:172.44ms
step:1306/1530 train_loss:3.4539 train_time:223496ms step_avg:172.45ms
step:1307/1530 train_loss:3.4555 train_time:223671ms step_avg:172.45ms
step:1308/1530 train_loss:3.2833 train_time:223851ms step_avg:172.46ms
step:1309/1530 train_loss:3.3106 train_time:224031ms step_avg:172.46ms
step:1310/1530 train_loss:3.2859 train_time:224209ms step_avg:172.47ms
step:1311/1530 train_loss:3.2964 train_time:224386ms step_avg:172.47ms
step:1312/1530 train_loss:3.3766 train_time:224566ms step_avg:172.48ms
step:1313/1530 train_loss:3.3397 train_time:224742ms step_avg:172.48ms
step:1314/1530 train_loss:3.0466 train_time:224926ms step_avg:172.49ms
step:1315/1530 train_loss:3.2740 train_time:225104ms step_avg:172.49ms
step:1316/1530 train_loss:3.3959 train_time:225281ms step_avg:172.50ms
step:1317/1530 train_loss:3.4211 train_time:225460ms step_avg:172.50ms
step:1318/1530 train_loss:3.3045 train_time:225647ms step_avg:172.51ms
step:1319/1530 train_loss:3.4282 train_time:225826ms step_avg:172.52ms
step:1320/1530 train_loss:3.4633 train_time:226009ms step_avg:172.53ms
step:1321/1530 train_loss:3.3625 train_time:226189ms step_avg:172.53ms
step:1322/1530 train_loss:3.3231 train_time:226503ms step_avg:172.64ms
step:1323/1530 train_loss:3.3211 train_time:226691ms step_avg:172.65ms
step:1324/1530 train_loss:3.4343 train_time:226871ms step_avg:172.66ms
step:1325/1530 train_loss:3.4903 train_time:227056ms step_avg:172.67ms
step:1326/1530 train_loss:3.2145 train_time:227237ms step_avg:172.67ms
step:1327/1530 train_loss:3.1653 train_time:227413ms step_avg:172.68ms
step:1328/1530 train_loss:3.4912 train_time:227592ms step_avg:172.68ms
step:1329/1530 train_loss:3.2948 train_time:227941ms step_avg:172.81ms
step:1330/1530 train_loss:3.4260 train_time:228123ms step_avg:172.82ms
step:1331/1530 train_loss:3.3315 train_time:228299ms step_avg:172.82ms
step:1332/1530 train_loss:3.7382 train_time:228480ms step_avg:172.83ms
step:1333/1530 train_loss:3.4797 train_time:228662ms step_avg:172.84ms
step:1334/1530 train_loss:3.3716 train_time:228841ms step_avg:172.84ms
step:1335/1530 train_loss:3.2907 train_time:229019ms step_avg:172.84ms
step:1336/1530 train_loss:3.2956 train_time:229205ms step_avg:172.85ms
step:1337/1530 train_loss:3.5478 train_time:229386ms step_avg:172.86ms
step:1338/1530 train_loss:3.5225 train_time:229565ms step_avg:172.87ms
step:1339/1530 train_loss:3.3407 train_time:229746ms step_avg:172.87ms
step:1340/1530 train_loss:3.2798 train_time:229924ms step_avg:172.87ms
step:1341/1530 train_loss:3.5934 train_time:230100ms step_avg:172.88ms
step:1342/1530 train_loss:3.3529 train_time:230280ms step_avg:172.88ms
step:1343/1530 train_loss:3.3641 train_time:230458ms step_avg:172.89ms
step:1344/1530 train_loss:3.4115 train_time:230639ms step_avg:172.89ms
step:1345/1530 train_loss:3.3833 train_time:230821ms step_avg:172.90ms
step:1346/1530 train_loss:3.3004 train_time:230997ms step_avg:172.90ms
step:1347/1530 train_loss:3.2795 train_time:231175ms step_avg:172.91ms
step:1348/1530 train_loss:3.3507 train_time:231354ms step_avg:172.91ms
step:1349/1530 train_loss:3.2749 train_time:231532ms step_avg:172.91ms
step:1350/1530 train_loss:3.3933 train_time:231712ms step_avg:172.92ms
step:1351/1530 train_loss:3.2425 train_time:231889ms step_avg:172.92ms
step:1352/1530 train_loss:3.3065 train_time:232067ms step_avg:172.93ms
step:1353/1530 train_loss:3.4037 train_time:232247ms step_avg:172.93ms
step:1354/1530 train_loss:3.2638 train_time:232423ms step_avg:172.93ms
step:1355/1530 train_loss:3.1882 train_time:232599ms step_avg:172.94ms
step:1356/1530 train_loss:3.5104 train_time:232779ms step_avg:172.94ms
step:1357/1530 train_loss:3.4250 train_time:232959ms step_avg:172.95ms
step:1358/1530 train_loss:3.1868 train_time:233139ms step_avg:172.95ms
step:1359/1530 train_loss:3.4424 train_time:233319ms step_avg:172.96ms
step:1360/1530 train_loss:3.3520 train_time:233499ms step_avg:172.96ms
step:1361/1530 train_loss:3.1191 train_time:233687ms step_avg:172.97ms
step:1362/1530 train_loss:3.3929 train_time:233867ms step_avg:172.98ms
step:1363/1530 train_loss:3.2850 train_time:234055ms step_avg:172.99ms
step:1364/1530 train_loss:3.3033 train_time:234233ms step_avg:172.99ms
step:1365/1530 train_loss:3.3161 train_time:234411ms step_avg:173.00ms
step:1366/1530 train_loss:3.4230 train_time:234592ms step_avg:173.00ms
step:1367/1530 train_loss:3.4004 train_time:234771ms step_avg:173.01ms
step:1368/1530 train_loss:3.3480 train_time:234951ms step_avg:173.01ms
step:1369/1530 train_loss:3.2752 train_time:235139ms step_avg:173.02ms
step:1370/1530 train_loss:3.6055 train_time:235319ms step_avg:173.03ms
step:1371/1530 train_loss:3.3153 train_time:235499ms step_avg:173.03ms
step:1372/1530 train_loss:3.3668 train_time:235681ms step_avg:173.04ms
step:1373/1530 train_loss:3.3703 train_time:235861ms step_avg:173.05ms
step:1374/1530 train_loss:3.1484 train_time:236043ms step_avg:173.05ms
step:1375/1530 train_loss:3.5374 train_time:236223ms step_avg:173.06ms
step:1375/1530 val_loss:3.3120 train_time:236274ms step_avg:173.09ms
step:1376/1530 train_loss:3.3502 train_time:236404ms step_avg:173.06ms
step:1377/1530 train_loss:3.4830 train_time:236584ms step_avg:173.07ms
step:1378/1530 train_loss:3.4750 train_time:236761ms step_avg:173.07ms
step:1379/1530 train_loss:3.1175 train_time:236944ms step_avg:173.08ms
step:1380/1530 train_loss:3.3130 train_time:237125ms step_avg:173.08ms
step:1381/1530 train_loss:3.7012 train_time:237310ms step_avg:173.09ms
step:1382/1530 train_loss:3.2100 train_time:237489ms step_avg:173.10ms
step:1383/1530 train_loss:3.3957 train_time:237670ms step_avg:173.10ms
step:1384/1530 train_loss:3.4759 train_time:237853ms step_avg:173.11ms
step:1385/1530 train_loss:3.4069 train_time:238028ms step_avg:173.11ms
step:1386/1530 train_loss:3.3488 train_time:238208ms step_avg:173.12ms
step:1387/1530 train_loss:3.2020 train_time:238387ms step_avg:173.12ms
step:1388/1530 train_loss:3.3442 train_time:238564ms step_avg:173.12ms
step:1389/1530 train_loss:3.3155 train_time:238749ms step_avg:173.13ms
step:1390/1530 train_loss:3.5679 train_time:238926ms step_avg:173.13ms
step:1391/1530 train_loss:3.2891 train_time:239104ms step_avg:173.14ms
step:1392/1530 train_loss:3.2899 train_time:239282ms step_avg:173.14ms
step:1393/1530 train_loss:3.2375 train_time:239466ms step_avg:173.15ms
step:1394/1530 train_loss:3.5028 train_time:239644ms step_avg:173.15ms
step:1395/1530 train_loss:3.3922 train_time:239822ms step_avg:173.16ms
step:1396/1530 train_loss:3.4064 train_time:239998ms step_avg:173.16ms
step:1397/1530 train_loss:3.3022 train_time:240174ms step_avg:173.16ms
step:1398/1530 train_loss:3.2542 train_time:240352ms step_avg:173.16ms
step:1399/1530 train_loss:3.3192 train_time:240533ms step_avg:173.17ms
step:1400/1530 train_loss:3.3225 train_time:240716ms step_avg:173.18ms
step:1401/1530 train_loss:3.3493 train_time:240891ms step_avg:173.18ms
step:1402/1530 train_loss:3.2976 train_time:241071ms step_avg:173.18ms
step:1403/1530 train_loss:3.4902 train_time:241255ms step_avg:173.19ms
step:1404/1530 train_loss:3.2809 train_time:241432ms step_avg:173.19ms
step:1405/1530 train_loss:3.3169 train_time:241613ms step_avg:173.20ms
step:1406/1530 train_loss:3.3150 train_time:241793ms step_avg:173.20ms
step:1407/1530 train_loss:3.1757 train_time:241970ms step_avg:173.21ms
step:1408/1530 train_loss:3.3144 train_time:242151ms step_avg:173.21ms
step:1409/1530 train_loss:3.3015 train_time:242338ms step_avg:173.22ms
step:1410/1530 train_loss:3.2896 train_time:242515ms step_avg:173.23ms
step:1411/1530 train_loss:3.3647 train_time:242690ms step_avg:173.23ms
step:1412/1530 train_loss:3.3341 train_time:242867ms step_avg:173.23ms
step:1413/1530 train_loss:3.3627 train_time:243045ms step_avg:173.23ms
step:1414/1530 train_loss:3.3323 train_time:243224ms step_avg:173.24ms
step:1415/1530 train_loss:3.4110 train_time:243408ms step_avg:173.24ms
step:1416/1530 train_loss:3.2319 train_time:243596ms step_avg:173.25ms
step:1417/1530 train_loss:3.2839 train_time:243779ms step_avg:173.26ms
step:1418/1530 train_loss:3.3872 train_time:243960ms step_avg:173.27ms
step:1419/1530 train_loss:3.3357 train_time:244144ms step_avg:173.27ms
step:1420/1530 train_loss:3.3636 train_time:244326ms step_avg:173.28ms
step:1421/1530 train_loss:3.3686 train_time:244506ms step_avg:173.29ms
step:1422/1530 train_loss:3.3304 train_time:244685ms step_avg:173.29ms
step:1423/1530 train_loss:3.3142 train_time:244864ms step_avg:173.29ms
step:1424/1530 train_loss:3.3359 train_time:245050ms step_avg:173.30ms
step:1425/1530 train_loss:3.1924 train_time:245236ms step_avg:173.31ms
step:1426/1530 train_loss:3.3239 train_time:245414ms step_avg:173.31ms
step:1427/1530 train_loss:3.2865 train_time:245597ms step_avg:173.32ms
step:1428/1530 train_loss:3.3768 train_time:245776ms step_avg:173.33ms
step:1429/1530 train_loss:3.3518 train_time:245953ms step_avg:173.33ms
step:1430/1530 train_loss:3.2607 train_time:246136ms step_avg:173.34ms
step:1431/1530 train_loss:3.3261 train_time:246318ms step_avg:173.34ms
step:1432/1530 train_loss:3.3341 train_time:246500ms step_avg:173.35ms
step:1433/1530 train_loss:3.1323 train_time:246684ms step_avg:173.36ms
step:1434/1530 train_loss:3.2884 train_time:246868ms step_avg:173.36ms
step:1435/1530 train_loss:3.1195 train_time:247048ms step_avg:173.37ms
step:1436/1530 train_loss:3.2317 train_time:247229ms step_avg:173.37ms
step:1437/1530 train_loss:3.4079 train_time:247406ms step_avg:173.38ms
step:1438/1530 train_loss:3.3843 train_time:247582ms step_avg:173.38ms
step:1439/1530 train_loss:3.3163 train_time:247761ms step_avg:173.38ms
step:1440/1530 train_loss:3.1912 train_time:247936ms step_avg:173.38ms
step:1441/1530 train_loss:3.3416 train_time:248114ms step_avg:173.39ms
step:1442/1530 train_loss:3.3858 train_time:248297ms step_avg:173.39ms
step:1443/1530 train_loss:3.4891 train_time:248483ms step_avg:173.40ms
step:1444/1530 train_loss:3.4484 train_time:248660ms step_avg:173.40ms
step:1445/1530 train_loss:3.3372 train_time:248838ms step_avg:173.41ms
step:1446/1530 train_loss:3.1971 train_time:249018ms step_avg:173.41ms
step:1447/1530 train_loss:3.2985 train_time:249200ms step_avg:173.42ms
step:1448/1530 train_loss:3.2926 train_time:249379ms step_avg:173.42ms
step:1449/1530 train_loss:3.3930 train_time:249558ms step_avg:173.42ms
step:1450/1530 train_loss:3.3836 train_time:249738ms step_avg:173.43ms
step:1451/1530 train_loss:3.2021 train_time:249915ms step_avg:173.43ms
step:1452/1530 train_loss:3.3264 train_time:250094ms step_avg:173.44ms
step:1453/1530 train_loss:3.2625 train_time:250269ms step_avg:173.44ms
step:1454/1530 train_loss:3.2885 train_time:250447ms step_avg:173.44ms
step:1455/1530 train_loss:3.3273 train_time:250630ms step_avg:173.45ms
step:1456/1530 train_loss:3.2849 train_time:250808ms step_avg:173.45ms
step:1457/1530 train_loss:3.1551 train_time:250984ms step_avg:173.45ms
step:1458/1530 train_loss:3.4237 train_time:251161ms step_avg:173.45ms
step:1459/1530 train_loss:3.2659 train_time:251343ms step_avg:173.46ms
step:1460/1530 train_loss:3.3167 train_time:251522ms step_avg:173.46ms
step:1461/1530 train_loss:3.4283 train_time:251703ms step_avg:173.47ms
step:1462/1530 train_loss:3.2635 train_time:251878ms step_avg:173.47ms
step:1463/1530 train_loss:3.4669 train_time:252062ms step_avg:173.48ms
step:1464/1530 train_loss:3.3645 train_time:252241ms step_avg:173.48ms
step:1465/1530 train_loss:3.3628 train_time:252421ms step_avg:173.49ms
step:1466/1530 train_loss:3.2851 train_time:252598ms step_avg:173.49ms
step:1467/1530 train_loss:3.3938 train_time:252779ms step_avg:173.49ms
step:1468/1530 train_loss:3.2891 train_time:252955ms step_avg:173.49ms
step:1469/1530 train_loss:3.2780 train_time:253134ms step_avg:173.50ms
step:1470/1530 train_loss:3.3325 train_time:253317ms step_avg:173.50ms
step:1471/1530 train_loss:3.2574 train_time:253501ms step_avg:173.51ms
step:1472/1530 train_loss:3.2454 train_time:253685ms step_avg:173.52ms
step:1473/1530 train_loss:3.4443 train_time:253863ms step_avg:173.52ms
step:1474/1530 train_loss:3.3141 train_time:254048ms step_avg:173.53ms
step:1475/1530 train_loss:3.1563 train_time:254233ms step_avg:173.54ms
step:1476/1530 train_loss:3.2696 train_time:254413ms step_avg:173.54ms
step:1477/1530 train_loss:3.2358 train_time:254600ms step_avg:173.55ms
step:1478/1530 train_loss:3.3082 train_time:254785ms step_avg:173.56ms
step:1479/1530 train_loss:3.3972 train_time:254965ms step_avg:173.56ms
step:1480/1530 train_loss:3.2673 train_time:255144ms step_avg:173.57ms
step:1481/1530 train_loss:3.4533 train_time:255327ms step_avg:173.57ms
step:1482/1530 train_loss:3.3669 train_time:255513ms step_avg:173.58ms
step:1483/1530 train_loss:3.2791 train_time:255705ms step_avg:173.59ms
step:1484/1530 train_loss:3.2655 train_time:255892ms step_avg:173.60ms
step:1485/1530 train_loss:3.2840 train_time:256072ms step_avg:173.61ms
step:1486/1530 train_loss:3.2244 train_time:256258ms step_avg:173.62ms
step:1487/1530 train_loss:3.3431 train_time:256440ms step_avg:173.62ms
step:1488/1530 train_loss:3.2437 train_time:256625ms step_avg:173.63ms
step:1489/1530 train_loss:3.3154 train_time:256805ms step_avg:173.63ms
step:1490/1530 train_loss:3.2531 train_time:256985ms step_avg:173.64ms
step:1491/1530 train_loss:3.1599 train_time:257167ms step_avg:173.64ms
step:1492/1530 train_loss:3.2693 train_time:257346ms step_avg:173.65ms
step:1493/1530 train_loss:3.4364 train_time:257527ms step_avg:173.65ms
step:1494/1530 train_loss:3.2940 train_time:257706ms step_avg:173.66ms
step:1495/1530 train_loss:3.0307 train_time:257890ms step_avg:173.66ms
step:1496/1530 train_loss:3.3579 train_time:258073ms step_avg:173.67ms
step:1497/1530 train_loss:3.3098 train_time:258256ms step_avg:173.68ms
step:1498/1530 train_loss:3.3515 train_time:258442ms step_avg:173.68ms
step:1499/1530 train_loss:3.3144 train_time:258629ms step_avg:173.69ms
step:1500/1530 train_loss:3.2963 train_time:258820ms step_avg:173.70ms
step:1500/1530 val_loss:3.2807 train_time:258875ms step_avg:173.74ms
step:1501/1530 train_loss:3.0869 train_time:259011ms step_avg:173.72ms
step:1502/1530 train_loss:3.3611 train_time:259202ms step_avg:173.73ms
step:1503/1530 train_loss:3.2427 train_time:259381ms step_avg:173.73ms
step:1504/1530 train_loss:3.2448 train_time:259564ms step_avg:173.74ms
step:1505/1530 train_loss:3.2099 train_time:259742ms step_avg:173.74ms
step:1506/1530 train_loss:3.2779 train_time:259925ms step_avg:173.75ms
step:1507/1530 train_loss:3.1812 train_time:260122ms step_avg:173.76ms
step:1508/1530 train_loss:3.4798 train_time:260304ms step_avg:173.77ms
step:1509/1530 train_loss:3.2801 train_time:260482ms step_avg:173.77ms
step:1510/1530 train_loss:3.2761 train_time:260662ms step_avg:173.77ms
step:1511/1530 train_loss:3.4136 train_time:260976ms step_avg:173.87ms
step:1512/1530 train_loss:3.4180 train_time:261161ms step_avg:173.88ms
step:1513/1530 train_loss:3.2704 train_time:261345ms step_avg:173.88ms
step:1514/1530 train_loss:3.0846 train_time:261529ms step_avg:173.89ms
step:1515/1530 train_loss:3.2473 train_time:261711ms step_avg:173.89ms
step:1516/1530 train_loss:3.2596 train_time:261895ms step_avg:173.90ms
step:1517/1530 train_loss:3.3000 train_time:262075ms step_avg:173.90ms
step:1518/1530 train_loss:3.2070 train_time:262258ms step_avg:173.91ms
step:1519/1530 train_loss:3.5052 train_time:262588ms step_avg:174.01ms
step:1520/1530 train_loss:3.1293 train_time:262773ms step_avg:174.02ms
step:1521/1530 train_loss:3.2034 train_time:262950ms step_avg:174.02ms
step:1522/1530 train_loss:3.3620 train_time:263135ms step_avg:174.03ms
step:1523/1530 train_loss:3.2340 train_time:263314ms step_avg:174.03ms
step:1524/1530 train_loss:3.3489 train_time:263497ms step_avg:174.04ms
step:1525/1530 train_loss:3.3351 train_time:263683ms step_avg:174.05ms
step:1526/1530 train_loss:3.2781 train_time:263874ms step_avg:174.06ms
step:1527/1530 train_loss:3.2945 train_time:264056ms step_avg:174.06ms
step:1528/1530 train_loss:3.4091 train_time:264236ms step_avg:174.07ms
step:1529/1530 train_loss:3.4099 train_time:264415ms step_avg:174.07ms
step:1530/1530 train_loss:3.2377 train_time:264593ms step_avg:174.07ms
step:1530/1530 val_loss:3.2783 train_time:264646ms step_avg:174.11ms