c9f224b2-0811-4d30-8fda-a28bd16a8ac0.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 01:54:25 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              96W / 700W |     25MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   2  NVIDIA H100 80GB HBM3          On  | 00000000:4C:00.0 Off |                    0 |
| N/A   31C    P0             117W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   3  NVIDIA H100 80GB HBM3          On  | 00000000:5D:00.0 Off |                    0 |
| N/A   38C    P0             119W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   4  NVIDIA H100 80GB HBM3          On  | 00000000:9B:00.0 Off |                    0 |
| N/A   39C    P0             123W / 700W |    529MiB / 81559MiB |      1%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   5  NVIDIA H100 80GB HBM3          On  | 00000000:BB:00.0 Off |                    0 |
| N/A   30C    P0             110W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   6  NVIDIA H100 80GB HBM3          On  | 00000000:CB:00.0 Off |                    0 |
| N/A   39C    P0             128W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   7  NVIDIA H100 80GB HBM3          On  | 00000000:DB:00.0 Off |                    0 |
| N/A   30C    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:31836ms step_avg:nanms
step:2/1530 train_loss:10.0830 train_time:31946ms step_avg:nanms
step:3/1530 train_loss:8.3903 train_time:32108ms step_avg:nanms
step:4/1530 train_loss:7.6005 train_time:32268ms step_avg:nanms
step:5/1530 train_loss:7.4460 train_time:32429ms step_avg:nanms
step:6/1530 train_loss:6.9756 train_time:32591ms step_avg:nanms
step:7/1530 train_loss:7.2050 train_time:32752ms step_avg:nanms
step:8/1530 train_loss:6.7317 train_time:32912ms step_avg:nanms
step:9/1530 train_loss:6.6337 train_time:33073ms step_avg:nanms
step:10/1530 train_loss:6.5364 train_time:33233ms step_avg:nanms
step:11/1530 train_loss:6.4976 train_time:115ms step_avg:nanms
step:12/1530 train_loss:6.3194 train_time:274ms step_avg:nanms
step:13/1530 train_loss:6.2342 train_time:434ms step_avg:144.81ms
step:14/1530 train_loss:6.1867 train_time:596ms step_avg:148.90ms
step:15/1530 train_loss:6.1904 train_time:756ms step_avg:151.17ms
step:16/1530 train_loss:6.1141 train_time:917ms step_avg:152.78ms
step:17/1530 train_loss:6.1608 train_time:1077ms step_avg:153.81ms
step:18/1530 train_loss:5.9522 train_time:1237ms step_avg:154.65ms
step:19/1530 train_loss:6.0026 train_time:1397ms step_avg:155.19ms
step:20/1530 train_loss:5.6724 train_time:1557ms step_avg:155.66ms
step:21/1530 train_loss:5.9656 train_time:1718ms step_avg:156.14ms
step:22/1530 train_loss:6.1818 train_time:1878ms step_avg:156.46ms
step:23/1530 train_loss:5.8547 train_time:2039ms step_avg:156.81ms
step:24/1530 train_loss:6.0110 train_time:2200ms step_avg:157.13ms
step:25/1530 train_loss:5.6919 train_time:2360ms step_avg:157.37ms
step:26/1530 train_loss:5.5946 train_time:2520ms step_avg:157.52ms
step:27/1530 train_loss:5.7919 train_time:2681ms step_avg:157.69ms
step:28/1530 train_loss:5.3969 train_time:2841ms step_avg:157.83ms
step:29/1530 train_loss:5.6730 train_time:3000ms step_avg:157.90ms
step:30/1530 train_loss:5.4515 train_time:3161ms step_avg:158.07ms
step:31/1530 train_loss:5.4338 train_time:3322ms step_avg:158.20ms
step:32/1530 train_loss:5.2804 train_time:3483ms step_avg:158.34ms
step:33/1530 train_loss:5.5872 train_time:3645ms step_avg:158.46ms
step:34/1530 train_loss:5.4997 train_time:3806ms step_avg:158.58ms
step:35/1530 train_loss:5.5985 train_time:3966ms step_avg:158.63ms
step:36/1530 train_loss:5.5329 train_time:4127ms step_avg:158.72ms
step:37/1530 train_loss:5.4617 train_time:4288ms step_avg:158.81ms
step:38/1530 train_loss:5.3164 train_time:4449ms step_avg:158.88ms
step:39/1530 train_loss:5.3390 train_time:4610ms step_avg:158.96ms
step:40/1530 train_loss:5.2389 train_time:4770ms step_avg:158.99ms
step:41/1530 train_loss:5.2283 train_time:4931ms step_avg:159.07ms
step:42/1530 train_loss:5.1635 train_time:5092ms step_avg:159.13ms
step:43/1530 train_loss:5.2701 train_time:5252ms step_avg:159.15ms
step:44/1530 train_loss:5.2293 train_time:5413ms step_avg:159.22ms
step:45/1530 train_loss:5.3798 train_time:5573ms step_avg:159.24ms
step:46/1530 train_loss:5.1579 train_time:5734ms step_avg:159.26ms
step:47/1530 train_loss:5.0793 train_time:5894ms step_avg:159.30ms
step:48/1530 train_loss:5.2072 train_time:6055ms step_avg:159.34ms
step:49/1530 train_loss:5.1477 train_time:6216ms step_avg:159.38ms
step:50/1530 train_loss:5.2458 train_time:6376ms step_avg:159.40ms
step:51/1530 train_loss:5.1230 train_time:6536ms step_avg:159.40ms
step:52/1530 train_loss:5.0260 train_time:6696ms step_avg:159.42ms
step:53/1530 train_loss:5.1872 train_time:6855ms step_avg:159.41ms
step:54/1530 train_loss:5.0179 train_time:7016ms step_avg:159.45ms
step:55/1530 train_loss:5.3976 train_time:7177ms step_avg:159.48ms
step:56/1530 train_loss:5.0104 train_time:7336ms step_avg:159.48ms
step:57/1530 train_loss:4.8742 train_time:7496ms step_avg:159.50ms
step:58/1530 train_loss:5.0458 train_time:7656ms step_avg:159.51ms
step:59/1530 train_loss:5.0422 train_time:7818ms step_avg:159.55ms
step:60/1530 train_loss:5.1648 train_time:7977ms step_avg:159.55ms
step:61/1530 train_loss:4.8575 train_time:8137ms step_avg:159.54ms
step:62/1530 train_loss:4.9853 train_time:8297ms step_avg:159.56ms
step:63/1530 train_loss:4.9839 train_time:8457ms step_avg:159.57ms
step:64/1530 train_loss:4.9733 train_time:8619ms step_avg:159.61ms
step:65/1530 train_loss:4.7947 train_time:8779ms step_avg:159.62ms
step:66/1530 train_loss:4.9088 train_time:8940ms step_avg:159.64ms
step:67/1530 train_loss:4.8210 train_time:9099ms step_avg:159.62ms
step:68/1530 train_loss:5.0910 train_time:9259ms step_avg:159.64ms
step:69/1530 train_loss:4.7207 train_time:9420ms step_avg:159.66ms
step:70/1530 train_loss:4.8223 train_time:9580ms step_avg:159.67ms
step:71/1530 train_loss:4.9670 train_time:9741ms step_avg:159.69ms
step:72/1530 train_loss:4.8837 train_time:9902ms step_avg:159.71ms
step:73/1530 train_loss:4.7690 train_time:10063ms step_avg:159.72ms
step:74/1530 train_loss:4.8931 train_time:10223ms step_avg:159.73ms
step:75/1530 train_loss:4.8518 train_time:10383ms step_avg:159.74ms
step:76/1530 train_loss:4.7910 train_time:10544ms step_avg:159.75ms
step:77/1530 train_loss:4.9127 train_time:10705ms step_avg:159.78ms
step:78/1530 train_loss:5.0956 train_time:10866ms step_avg:159.79ms
step:79/1530 train_loss:4.8328 train_time:11026ms step_avg:159.80ms
step:80/1530 train_loss:4.8496 train_time:11187ms step_avg:159.81ms
step:81/1530 train_loss:4.6508 train_time:11347ms step_avg:159.82ms
step:82/1530 train_loss:4.8118 train_time:11508ms step_avg:159.83ms
step:83/1530 train_loss:4.7833 train_time:11668ms step_avg:159.83ms
step:84/1530 train_loss:4.7816 train_time:11829ms step_avg:159.85ms
step:85/1530 train_loss:4.6173 train_time:11989ms step_avg:159.86ms
step:86/1530 train_loss:4.8352 train_time:12150ms step_avg:159.87ms
step:87/1530 train_loss:4.7510 train_time:12311ms step_avg:159.88ms
step:88/1530 train_loss:4.7364 train_time:12471ms step_avg:159.89ms
step:89/1530 train_loss:4.6888 train_time:12632ms step_avg:159.90ms
step:90/1530 train_loss:4.6390 train_time:12792ms step_avg:159.90ms
step:91/1530 train_loss:4.6292 train_time:12952ms step_avg:159.90ms
step:92/1530 train_loss:4.7862 train_time:13114ms step_avg:159.92ms
step:93/1530 train_loss:4.6221 train_time:13274ms step_avg:159.93ms
step:94/1530 train_loss:4.6515 train_time:13434ms step_avg:159.93ms
step:95/1530 train_loss:4.7000 train_time:13595ms step_avg:159.94ms
step:96/1530 train_loss:4.5891 train_time:13755ms step_avg:159.94ms
step:97/1530 train_loss:4.6517 train_time:13916ms step_avg:159.95ms
step:98/1530 train_loss:4.5980 train_time:14076ms step_avg:159.96ms
step:99/1530 train_loss:4.6764 train_time:14236ms step_avg:159.95ms
step:100/1530 train_loss:4.6909 train_time:14397ms step_avg:159.97ms
step:101/1530 train_loss:4.5762 train_time:14557ms step_avg:159.96ms
step:102/1530 train_loss:4.7075 train_time:14718ms step_avg:159.98ms
step:103/1530 train_loss:4.5816 train_time:14878ms step_avg:159.98ms
step:104/1530 train_loss:4.5371 train_time:15038ms step_avg:159.98ms
step:105/1530 train_loss:4.5531 train_time:15198ms step_avg:159.98ms
step:106/1530 train_loss:4.6095 train_time:15358ms step_avg:159.98ms
step:107/1530 train_loss:4.5079 train_time:15520ms step_avg:160.00ms
step:108/1530 train_loss:4.3609 train_time:15681ms step_avg:160.01ms
step:109/1530 train_loss:4.4814 train_time:15842ms step_avg:160.02ms
step:110/1530 train_loss:4.4816 train_time:16002ms step_avg:160.02ms
step:111/1530 train_loss:4.4195 train_time:16164ms step_avg:160.04ms
step:112/1530 train_loss:4.6003 train_time:16325ms step_avg:160.05ms
step:113/1530 train_loss:4.5017 train_time:16486ms step_avg:160.06ms
step:114/1530 train_loss:4.3657 train_time:16646ms step_avg:160.06ms
step:115/1530 train_loss:4.5124 train_time:16809ms step_avg:160.08ms
step:116/1530 train_loss:4.4749 train_time:16972ms step_avg:160.11ms
step:117/1530 train_loss:4.3761 train_time:17136ms step_avg:160.15ms
step:118/1530 train_loss:4.5964 train_time:17301ms step_avg:160.20ms
step:119/1530 train_loss:4.4608 train_time:17466ms step_avg:160.24ms
step:120/1530 train_loss:4.3305 train_time:17630ms step_avg:160.28ms
step:121/1530 train_loss:4.2997 train_time:17794ms step_avg:160.31ms
step:122/1530 train_loss:4.4482 train_time:17958ms step_avg:160.34ms
step:123/1530 train_loss:4.2917 train_time:18122ms step_avg:160.37ms
step:124/1530 train_loss:4.6052 train_time:18286ms step_avg:160.41ms
step:125/1530 train_loss:4.4678 train_time:18450ms step_avg:160.44ms
step:125/1530 val_loss:4.4138 train_time:18497ms step_avg:160.85ms
step:126/1530 train_loss:4.4291 train_time:18617ms step_avg:160.49ms
step:127/1530 train_loss:4.4439 train_time:18782ms step_avg:160.53ms
step:128/1530 train_loss:4.3688 train_time:18947ms step_avg:160.57ms
step:129/1530 train_loss:4.6847 train_time:19111ms step_avg:160.60ms
step:130/1530 train_loss:4.3691 train_time:19275ms step_avg:160.63ms
step:131/1530 train_loss:4.4033 train_time:19439ms step_avg:160.65ms
step:132/1530 train_loss:4.3454 train_time:19603ms step_avg:160.68ms
step:133/1530 train_loss:4.4549 train_time:19767ms step_avg:160.70ms
step:134/1530 train_loss:4.2588 train_time:19931ms step_avg:160.73ms
step:135/1530 train_loss:4.4517 train_time:20096ms step_avg:160.77ms
step:136/1530 train_loss:4.2133 train_time:20259ms step_avg:160.79ms
step:137/1530 train_loss:4.3848 train_time:20423ms step_avg:160.81ms
step:138/1530 train_loss:4.2782 train_time:20587ms step_avg:160.84ms
step:139/1530 train_loss:4.3801 train_time:20750ms step_avg:160.85ms
step:140/1530 train_loss:4.4927 train_time:20914ms step_avg:160.88ms
step:141/1530 train_loss:4.3319 train_time:21078ms step_avg:160.90ms
step:142/1530 train_loss:4.3184 train_time:21241ms step_avg:160.92ms
step:143/1530 train_loss:4.2610 train_time:21405ms step_avg:160.94ms
step:144/1530 train_loss:4.3539 train_time:21570ms step_avg:160.97ms
step:145/1530 train_loss:4.3100 train_time:21734ms step_avg:160.99ms
step:146/1530 train_loss:4.1749 train_time:21898ms step_avg:161.01ms
step:147/1530 train_loss:4.3273 train_time:22062ms step_avg:161.03ms
step:148/1530 train_loss:4.3682 train_time:22224ms step_avg:161.04ms
step:149/1530 train_loss:4.3118 train_time:22390ms step_avg:161.08ms
step:150/1530 train_loss:4.4525 train_time:22552ms step_avg:161.09ms
step:151/1530 train_loss:4.2802 train_time:22716ms step_avg:161.11ms
step:152/1530 train_loss:4.2780 train_time:22880ms step_avg:161.13ms
step:153/1530 train_loss:4.3665 train_time:23044ms step_avg:161.15ms
step:154/1530 train_loss:4.3764 train_time:23208ms step_avg:161.17ms
step:155/1530 train_loss:4.2873 train_time:23373ms step_avg:161.19ms
step:156/1530 train_loss:4.3577 train_time:23536ms step_avg:161.20ms
step:157/1530 train_loss:4.4157 train_time:23701ms step_avg:161.23ms
step:158/1530 train_loss:4.2578 train_time:23864ms step_avg:161.25ms
step:159/1530 train_loss:4.3145 train_time:24028ms step_avg:161.26ms
step:160/1530 train_loss:4.1492 train_time:24192ms step_avg:161.28ms
step:161/1530 train_loss:4.3548 train_time:24356ms step_avg:161.30ms
step:162/1530 train_loss:4.3687 train_time:24519ms step_avg:161.31ms
step:163/1530 train_loss:4.3435 train_time:24684ms step_avg:161.33ms
step:164/1530 train_loss:4.1943 train_time:24847ms step_avg:161.35ms
step:165/1530 train_loss:4.2867 train_time:25010ms step_avg:161.36ms
step:166/1530 train_loss:4.3411 train_time:25176ms step_avg:161.38ms
step:167/1530 train_loss:4.2098 train_time:25339ms step_avg:161.39ms
step:168/1530 train_loss:4.3012 train_time:25503ms step_avg:161.41ms
step:169/1530 train_loss:4.1672 train_time:25668ms step_avg:161.43ms
step:170/1530 train_loss:4.0301 train_time:25833ms step_avg:161.46ms
step:171/1530 train_loss:4.2054 train_time:25997ms step_avg:161.47ms
step:172/1530 train_loss:4.2169 train_time:26160ms step_avg:161.48ms
step:173/1530 train_loss:4.2664 train_time:26324ms step_avg:161.49ms
step:174/1530 train_loss:4.4242 train_time:26486ms step_avg:161.50ms
step:175/1530 train_loss:4.2458 train_time:26650ms step_avg:161.51ms
step:176/1530 train_loss:4.0964 train_time:26813ms step_avg:161.52ms
step:177/1530 train_loss:4.0688 train_time:26976ms step_avg:161.53ms
step:178/1530 train_loss:4.1924 train_time:27137ms step_avg:161.53ms
step:179/1530 train_loss:4.1350 train_time:27301ms step_avg:161.55ms
step:180/1530 train_loss:4.1211 train_time:27464ms step_avg:161.55ms
step:181/1530 train_loss:4.3001 train_time:27627ms step_avg:161.56ms
step:182/1530 train_loss:4.1543 train_time:27791ms step_avg:161.57ms
step:183/1530 train_loss:4.1299 train_time:27954ms step_avg:161.58ms
step:184/1530 train_loss:4.1257 train_time:28116ms step_avg:161.59ms
step:185/1530 train_loss:4.2148 train_time:28279ms step_avg:161.59ms
step:186/1530 train_loss:4.1749 train_time:28442ms step_avg:161.60ms
step:187/1530 train_loss:4.2371 train_time:28606ms step_avg:161.61ms
step:188/1530 train_loss:4.1724 train_time:28900ms step_avg:162.36ms
step:189/1530 train_loss:4.1052 train_time:29225ms step_avg:163.27ms
step:190/1530 train_loss:4.2076 train_time:29388ms step_avg:163.27ms
step:191/1530 train_loss:4.0863 train_time:29551ms step_avg:163.27ms
step:192/1530 train_loss:4.0382 train_time:29714ms step_avg:163.26ms
step:193/1530 train_loss:4.2555 train_time:29878ms step_avg:163.27ms
step:194/1530 train_loss:4.1748 train_time:30039ms step_avg:163.26ms
step:195/1530 train_loss:4.3538 train_time:30203ms step_avg:163.26ms
step:196/1530 train_loss:4.1774 train_time:30366ms step_avg:163.26ms
step:197/1530 train_loss:4.0503 train_time:30530ms step_avg:163.26ms
step:198/1530 train_loss:4.1873 train_time:30692ms step_avg:163.26ms
step:199/1530 train_loss:4.0379 train_time:30856ms step_avg:163.26ms
step:200/1530 train_loss:4.1120 train_time:31020ms step_avg:163.26ms
step:201/1530 train_loss:4.0305 train_time:31184ms step_avg:163.27ms
step:202/1530 train_loss:4.2808 train_time:31347ms step_avg:163.27ms
step:203/1530 train_loss:4.0678 train_time:31510ms step_avg:163.26ms
step:204/1530 train_loss:4.1951 train_time:31672ms step_avg:163.26ms
step:205/1530 train_loss:4.2515 train_time:31835ms step_avg:163.26ms
step:206/1530 train_loss:3.9472 train_time:31998ms step_avg:163.26ms
step:207/1530 train_loss:4.0882 train_time:32161ms step_avg:163.25ms
step:208/1530 train_loss:4.1040 train_time:32324ms step_avg:163.25ms
step:209/1530 train_loss:4.2415 train_time:32488ms step_avg:163.26ms
step:210/1530 train_loss:4.1770 train_time:32650ms step_avg:163.25ms
step:211/1530 train_loss:4.0642 train_time:32812ms step_avg:163.25ms
step:212/1530 train_loss:4.1245 train_time:32976ms step_avg:163.25ms
step:213/1530 train_loss:4.0552 train_time:33138ms step_avg:163.24ms
step:214/1530 train_loss:4.1187 train_time:33302ms step_avg:163.25ms
step:215/1530 train_loss:3.9807 train_time:33466ms step_avg:163.25ms
step:216/1530 train_loss:4.0037 train_time:33629ms step_avg:163.25ms
step:217/1530 train_loss:4.0155 train_time:33792ms step_avg:163.25ms
step:218/1530 train_loss:4.0832 train_time:33955ms step_avg:163.24ms
step:219/1530 train_loss:4.0803 train_time:34117ms step_avg:163.24ms
step:220/1530 train_loss:4.0895 train_time:34280ms step_avg:163.24ms
step:221/1530 train_loss:4.0951 train_time:34443ms step_avg:163.24ms
step:222/1530 train_loss:4.0028 train_time:34605ms step_avg:163.23ms
step:223/1530 train_loss:3.9958 train_time:34768ms step_avg:163.23ms
step:224/1530 train_loss:4.3052 train_time:34932ms step_avg:163.23ms
step:225/1530 train_loss:3.9425 train_time:35094ms step_avg:163.23ms
step:226/1530 train_loss:3.9916 train_time:35258ms step_avg:163.23ms
step:227/1530 train_loss:3.9800 train_time:35421ms step_avg:163.23ms
step:228/1530 train_loss:4.1571 train_time:35586ms step_avg:163.24ms
step:229/1530 train_loss:3.9267 train_time:35754ms step_avg:163.26ms
step:230/1530 train_loss:4.0470 train_time:35919ms step_avg:163.27ms
step:231/1530 train_loss:3.9007 train_time:36085ms step_avg:163.28ms
step:232/1530 train_loss:3.9724 train_time:36252ms step_avg:163.30ms
step:233/1530 train_loss:4.1012 train_time:36418ms step_avg:163.31ms
step:234/1530 train_loss:4.0392 train_time:36584ms step_avg:163.32ms
step:235/1530 train_loss:3.9162 train_time:36751ms step_avg:163.34ms
step:236/1530 train_loss:4.0889 train_time:36917ms step_avg:163.35ms
step:237/1530 train_loss:4.0886 train_time:37083ms step_avg:163.36ms
step:238/1530 train_loss:3.9440 train_time:37251ms step_avg:163.38ms
step:239/1530 train_loss:4.0757 train_time:37416ms step_avg:163.39ms
step:240/1530 train_loss:4.1178 train_time:37582ms step_avg:163.40ms
step:241/1530 train_loss:3.9771 train_time:37748ms step_avg:163.41ms
step:242/1530 train_loss:4.1532 train_time:37915ms step_avg:163.43ms
step:243/1530 train_loss:4.0153 train_time:38081ms step_avg:163.44ms
step:244/1530 train_loss:4.0835 train_time:38246ms step_avg:163.44ms
step:245/1530 train_loss:4.1447 train_time:38412ms step_avg:163.46ms
step:246/1530 train_loss:4.0565 train_time:38578ms step_avg:163.47ms
step:247/1530 train_loss:4.0011 train_time:38743ms step_avg:163.47ms
step:248/1530 train_loss:4.1098 train_time:38909ms step_avg:163.49ms
step:249/1530 train_loss:3.9350 train_time:39076ms step_avg:163.50ms
step:250/1530 train_loss:3.9870 train_time:39240ms step_avg:163.50ms
step:250/1530 val_loss:4.0128 train_time:39289ms step_avg:163.70ms
step:251/1530 train_loss:4.0836 train_time:39407ms step_avg:163.51ms
step:252/1530 train_loss:4.1638 train_time:39573ms step_avg:163.53ms
step:253/1530 train_loss:3.9297 train_time:39740ms step_avg:163.54ms
step:254/1530 train_loss:3.8841 train_time:39906ms step_avg:163.55ms
step:255/1530 train_loss:4.0874 train_time:40073ms step_avg:163.56ms
step:256/1530 train_loss:3.9898 train_time:40239ms step_avg:163.57ms
step:257/1530 train_loss:3.9890 train_time:40404ms step_avg:163.58ms
step:258/1530 train_loss:3.9823 train_time:40570ms step_avg:163.59ms
step:259/1530 train_loss:4.0285 train_time:40737ms step_avg:163.60ms
step:260/1530 train_loss:4.0626 train_time:40903ms step_avg:163.61ms
step:261/1530 train_loss:4.0279 train_time:41069ms step_avg:163.62ms
step:262/1530 train_loss:4.0034 train_time:41236ms step_avg:163.63ms
step:263/1530 train_loss:3.8995 train_time:41401ms step_avg:163.64ms
step:264/1530 train_loss:3.9932 train_time:41567ms step_avg:163.65ms
step:265/1530 train_loss:3.8704 train_time:41735ms step_avg:163.67ms
step:266/1530 train_loss:3.9227 train_time:41900ms step_avg:163.67ms
step:267/1530 train_loss:3.9264 train_time:42066ms step_avg:163.68ms
step:268/1530 train_loss:3.9635 train_time:42233ms step_avg:163.69ms
step:269/1530 train_loss:3.8653 train_time:42398ms step_avg:163.70ms
step:270/1530 train_loss:4.1047 train_time:42564ms step_avg:163.71ms
step:271/1530 train_loss:3.9693 train_time:42732ms step_avg:163.72ms
step:272/1530 train_loss:3.9309 train_time:42897ms step_avg:163.73ms
step:273/1530 train_loss:3.9515 train_time:43063ms step_avg:163.74ms
step:274/1530 train_loss:4.0429 train_time:43230ms step_avg:163.75ms
step:275/1530 train_loss:4.0610 train_time:43395ms step_avg:163.75ms
step:276/1530 train_loss:4.2261 train_time:43562ms step_avg:163.77ms
step:277/1530 train_loss:4.0411 train_time:43728ms step_avg:163.77ms
step:278/1530 train_loss:4.0941 train_time:43895ms step_avg:163.79ms
step:279/1530 train_loss:4.0076 train_time:44061ms step_avg:163.80ms
step:280/1530 train_loss:4.1876 train_time:44230ms step_avg:163.81ms
step:281/1530 train_loss:3.9810 train_time:44396ms step_avg:163.82ms
step:282/1530 train_loss:3.9548 train_time:44562ms step_avg:163.83ms
step:283/1530 train_loss:3.9184 train_time:44729ms step_avg:163.84ms
step:284/1530 train_loss:4.0543 train_time:44894ms step_avg:163.85ms
step:285/1530 train_loss:4.0687 train_time:45059ms step_avg:163.85ms
step:286/1530 train_loss:4.0927 train_time:45224ms step_avg:163.86ms
step:287/1530 train_loss:3.9132 train_time:45389ms step_avg:163.86ms
step:288/1530 train_loss:4.0175 train_time:45554ms step_avg:163.86ms
step:289/1530 train_loss:3.8779 train_time:45719ms step_avg:163.87ms
step:290/1530 train_loss:3.8589 train_time:45885ms step_avg:163.88ms
step:291/1530 train_loss:3.9129 train_time:46053ms step_avg:163.89ms
step:292/1530 train_loss:3.8699 train_time:46218ms step_avg:163.89ms
step:293/1530 train_loss:3.9122 train_time:46383ms step_avg:163.90ms
step:294/1530 train_loss:3.9351 train_time:46549ms step_avg:163.90ms
step:295/1530 train_loss:3.8432 train_time:46715ms step_avg:163.91ms
step:296/1530 train_loss:3.8649 train_time:46880ms step_avg:163.91ms
step:297/1530 train_loss:3.8661 train_time:47047ms step_avg:163.93ms
step:298/1530 train_loss:3.9737 train_time:47214ms step_avg:163.94ms
step:299/1530 train_loss:3.8231 train_time:47378ms step_avg:163.94ms
step:300/1530 train_loss:3.9705 train_time:47544ms step_avg:163.95ms
step:301/1530 train_loss:3.9615 train_time:47710ms step_avg:163.95ms
step:302/1530 train_loss:3.9311 train_time:47875ms step_avg:163.95ms
step:303/1530 train_loss:3.9812 train_time:48039ms step_avg:163.96ms
step:304/1530 train_loss:3.9800 train_time:48204ms step_avg:163.96ms
step:305/1530 train_loss:4.4519 train_time:48370ms step_avg:163.97ms
step:306/1530 train_loss:3.9402 train_time:48535ms step_avg:163.97ms
step:307/1530 train_loss:3.8364 train_time:48699ms step_avg:163.97ms
step:308/1530 train_loss:3.9809 train_time:48865ms step_avg:163.98ms
step:309/1530 train_loss:3.8679 train_time:49032ms step_avg:163.99ms
step:310/1530 train_loss:4.0873 train_time:49196ms step_avg:163.99ms
step:311/1530 train_loss:3.9313 train_time:49362ms step_avg:163.99ms
step:312/1530 train_loss:3.8664 train_time:49527ms step_avg:164.00ms
step:313/1530 train_loss:3.9420 train_time:49693ms step_avg:164.00ms
step:314/1530 train_loss:4.0698 train_time:49858ms step_avg:164.01ms
step:315/1530 train_loss:3.9500 train_time:50022ms step_avg:164.01ms
step:316/1530 train_loss:3.7981 train_time:50189ms step_avg:164.02ms
step:317/1530 train_loss:3.8775 train_time:50355ms step_avg:164.02ms
step:318/1530 train_loss:3.9296 train_time:50520ms step_avg:164.03ms
step:319/1530 train_loss:3.8957 train_time:50685ms step_avg:164.03ms
step:320/1530 train_loss:4.0125 train_time:50851ms step_avg:164.04ms
step:321/1530 train_loss:3.9669 train_time:51016ms step_avg:164.04ms
step:322/1530 train_loss:3.9346 train_time:51181ms step_avg:164.04ms
step:323/1530 train_loss:4.0057 train_time:51347ms step_avg:164.05ms
step:324/1530 train_loss:3.9557 train_time:51513ms step_avg:164.05ms
step:325/1530 train_loss:4.0188 train_time:51678ms step_avg:164.06ms
step:326/1530 train_loss:3.9010 train_time:51843ms step_avg:164.06ms
step:327/1530 train_loss:4.3986 train_time:52009ms step_avg:164.07ms
step:328/1530 train_loss:4.0737 train_time:52176ms step_avg:164.07ms
step:329/1530 train_loss:3.7895 train_time:52341ms step_avg:164.08ms
step:330/1530 train_loss:3.7523 train_time:52506ms step_avg:164.08ms
step:331/1530 train_loss:3.9770 train_time:52671ms step_avg:164.09ms
step:332/1530 train_loss:3.9194 train_time:52836ms step_avg:164.09ms
step:333/1530 train_loss:3.9002 train_time:53001ms step_avg:164.09ms
step:334/1530 train_loss:3.8522 train_time:53166ms step_avg:164.09ms
step:335/1530 train_loss:4.0170 train_time:53332ms step_avg:164.10ms
step:336/1530 train_loss:3.9633 train_time:53496ms step_avg:164.10ms
step:337/1530 train_loss:4.4253 train_time:53662ms step_avg:164.10ms
step:338/1530 train_loss:3.9450 train_time:53828ms step_avg:164.11ms
step:339/1530 train_loss:3.8684 train_time:53993ms step_avg:164.11ms
step:340/1530 train_loss:3.9442 train_time:54158ms step_avg:164.11ms
step:341/1530 train_loss:3.8598 train_time:54325ms step_avg:164.12ms
step:342/1530 train_loss:3.8184 train_time:54493ms step_avg:164.14ms
step:343/1530 train_loss:3.8419 train_time:54661ms step_avg:164.15ms
step:344/1530 train_loss:4.0002 train_time:54828ms step_avg:164.16ms
step:345/1530 train_loss:3.8253 train_time:54996ms step_avg:164.17ms
step:346/1530 train_loss:3.7738 train_time:55165ms step_avg:164.18ms
step:347/1530 train_loss:3.8071 train_time:55337ms step_avg:164.20ms
step:348/1530 train_loss:3.8655 train_time:55504ms step_avg:164.21ms
step:349/1530 train_loss:3.8293 train_time:55672ms step_avg:164.22ms
step:350/1530 train_loss:3.5644 train_time:55841ms step_avg:164.24ms
step:351/1530 train_loss:3.8261 train_time:56008ms step_avg:164.25ms
step:352/1530 train_loss:4.1770 train_time:56177ms step_avg:164.26ms
step:353/1530 train_loss:3.6627 train_time:56345ms step_avg:164.27ms
step:354/1530 train_loss:3.9311 train_time:56513ms step_avg:164.28ms
step:355/1530 train_loss:3.7873 train_time:56681ms step_avg:164.29ms
step:356/1530 train_loss:3.8887 train_time:56852ms step_avg:164.31ms
step:357/1530 train_loss:3.7738 train_time:57021ms step_avg:164.32ms
step:358/1530 train_loss:3.8668 train_time:57189ms step_avg:164.34ms
step:359/1530 train_loss:3.7766 train_time:57360ms step_avg:164.35ms
step:360/1530 train_loss:3.4336 train_time:57529ms step_avg:164.37ms
step:361/1530 train_loss:4.0285 train_time:57697ms step_avg:164.38ms
step:362/1530 train_loss:3.9250 train_time:57864ms step_avg:164.39ms
step:363/1530 train_loss:3.8428 train_time:58034ms step_avg:164.40ms
step:364/1530 train_loss:3.7492 train_time:58203ms step_avg:164.42ms
step:365/1530 train_loss:3.9145 train_time:58373ms step_avg:164.43ms
step:366/1530 train_loss:3.8651 train_time:58541ms step_avg:164.44ms
step:367/1530 train_loss:3.8656 train_time:58708ms step_avg:164.45ms
step:368/1530 train_loss:3.8568 train_time:58876ms step_avg:164.46ms
step:369/1530 train_loss:3.7493 train_time:59045ms step_avg:164.47ms
step:370/1530 train_loss:3.8819 train_time:59213ms step_avg:164.48ms
step:371/1530 train_loss:3.7332 train_time:59381ms step_avg:164.49ms
step:372/1530 train_loss:3.6901 train_time:59549ms step_avg:164.50ms
step:373/1530 train_loss:3.9117 train_time:59716ms step_avg:164.51ms
step:374/1530 train_loss:3.8243 train_time:59884ms step_avg:164.52ms
step:375/1530 train_loss:3.8034 train_time:60053ms step_avg:164.53ms
step:375/1530 val_loss:3.8290 train_time:60100ms step_avg:164.66ms
step:376/1530 train_loss:3.8722 train_time:60220ms step_avg:164.54ms
step:377/1530 train_loss:3.7956 train_time:60520ms step_avg:164.90ms
step:378/1530 train_loss:3.8589 train_time:60698ms step_avg:164.94ms
step:379/1530 train_loss:3.8775 train_time:61016ms step_avg:165.36ms
step:380/1530 train_loss:3.9478 train_time:61185ms step_avg:165.37ms
step:381/1530 train_loss:3.8405 train_time:61352ms step_avg:165.37ms
step:382/1530 train_loss:3.8095 train_time:61521ms step_avg:165.38ms
step:383/1530 train_loss:3.8013 train_time:61691ms step_avg:165.39ms
step:384/1530 train_loss:3.8767 train_time:61859ms step_avg:165.40ms
step:385/1530 train_loss:3.7908 train_time:62027ms step_avg:165.41ms
step:386/1530 train_loss:3.8953 train_time:62195ms step_avg:165.41ms
step:387/1530 train_loss:4.0584 train_time:62364ms step_avg:165.42ms
step:388/1530 train_loss:3.7992 train_time:62532ms step_avg:165.43ms
step:389/1530 train_loss:3.7999 train_time:62700ms step_avg:165.43ms
step:390/1530 train_loss:3.9060 train_time:62869ms step_avg:165.44ms
step:391/1530 train_loss:3.8172 train_time:63037ms step_avg:165.45ms
step:392/1530 train_loss:3.9313 train_time:63206ms step_avg:165.46ms
step:393/1530 train_loss:3.7728 train_time:63372ms step_avg:165.46ms
step:394/1530 train_loss:3.8944 train_time:63540ms step_avg:165.47ms
step:395/1530 train_loss:3.6326 train_time:63708ms step_avg:165.48ms
step:396/1530 train_loss:3.8422 train_time:63876ms step_avg:165.48ms
step:397/1530 train_loss:3.8680 train_time:64043ms step_avg:165.49ms
step:398/1530 train_loss:3.8896 train_time:64211ms step_avg:165.49ms
step:399/1530 train_loss:3.7719 train_time:64377ms step_avg:165.49ms
step:400/1530 train_loss:3.8208 train_time:64546ms step_avg:165.50ms
step:401/1530 train_loss:3.9198 train_time:64713ms step_avg:165.51ms
step:402/1530 train_loss:3.8444 train_time:64880ms step_avg:165.51ms
step:403/1530 train_loss:3.9674 train_time:65048ms step_avg:165.52ms
step:404/1530 train_loss:3.6814 train_time:65215ms step_avg:165.52ms
step:405/1530 train_loss:3.7886 train_time:65383ms step_avg:165.53ms
step:406/1530 train_loss:4.0958 train_time:65550ms step_avg:165.53ms
step:407/1530 train_loss:3.7757 train_time:65716ms step_avg:165.53ms
step:408/1530 train_loss:3.8232 train_time:65884ms step_avg:165.54ms
step:409/1530 train_loss:3.8575 train_time:66051ms step_avg:165.54ms
step:410/1530 train_loss:3.7617 train_time:66218ms step_avg:165.54ms
step:411/1530 train_loss:3.7686 train_time:66386ms step_avg:165.55ms
step:412/1530 train_loss:4.1910 train_time:66553ms step_avg:165.55ms
step:413/1530 train_loss:3.6626 train_time:66719ms step_avg:165.56ms
step:414/1530 train_loss:4.0118 train_time:66888ms step_avg:165.56ms
step:415/1530 train_loss:3.7554 train_time:67054ms step_avg:165.56ms
step:416/1530 train_loss:3.7689 train_time:67221ms step_avg:165.57ms
step:417/1530 train_loss:3.9587 train_time:67389ms step_avg:165.58ms
step:418/1530 train_loss:3.6966 train_time:67555ms step_avg:165.58ms
step:419/1530 train_loss:3.8069 train_time:67722ms step_avg:165.58ms
step:420/1530 train_loss:3.7041 train_time:67890ms step_avg:165.58ms
step:421/1530 train_loss:3.6522 train_time:68057ms step_avg:165.59ms
step:422/1530 train_loss:3.7895 train_time:68225ms step_avg:165.60ms
step:423/1530 train_loss:3.8788 train_time:68392ms step_avg:165.60ms
step:424/1530 train_loss:3.6119 train_time:68560ms step_avg:165.60ms
step:425/1530 train_loss:3.8015 train_time:68727ms step_avg:165.61ms
step:426/1530 train_loss:3.6628 train_time:68895ms step_avg:165.61ms
step:427/1530 train_loss:3.8933 train_time:69062ms step_avg:165.62ms
step:428/1530 train_loss:3.8179 train_time:69230ms step_avg:165.62ms
step:429/1530 train_loss:3.7653 train_time:69397ms step_avg:165.63ms
step:430/1530 train_loss:3.7060 train_time:69566ms step_avg:165.63ms
step:431/1530 train_loss:3.6297 train_time:69734ms step_avg:165.64ms
step:432/1530 train_loss:3.7648 train_time:69904ms step_avg:165.65ms
step:433/1530 train_loss:3.8239 train_time:70070ms step_avg:165.65ms
step:434/1530 train_loss:3.7818 train_time:70237ms step_avg:165.65ms
step:435/1530 train_loss:3.8214 train_time:70404ms step_avg:165.66ms
step:436/1530 train_loss:3.8324 train_time:70572ms step_avg:165.66ms
step:437/1530 train_loss:3.7187 train_time:70739ms step_avg:165.67ms
step:438/1530 train_loss:3.7041 train_time:70908ms step_avg:165.67ms
step:439/1530 train_loss:3.7173 train_time:71075ms step_avg:165.68ms
step:440/1530 train_loss:3.8923 train_time:71243ms step_avg:165.68ms
step:441/1530 train_loss:3.7609 train_time:71411ms step_avg:165.69ms
step:442/1530 train_loss:3.7407 train_time:71577ms step_avg:165.69ms
step:443/1530 train_loss:3.6273 train_time:71745ms step_avg:165.69ms
step:444/1530 train_loss:3.9178 train_time:71912ms step_avg:165.69ms
step:445/1530 train_loss:3.8446 train_time:72078ms step_avg:165.70ms
step:446/1530 train_loss:3.8415 train_time:72246ms step_avg:165.70ms
step:447/1530 train_loss:3.7550 train_time:72412ms step_avg:165.70ms
step:448/1530 train_loss:3.8537 train_time:72578ms step_avg:165.70ms
step:449/1530 train_loss:3.6901 train_time:72747ms step_avg:165.71ms
step:450/1530 train_loss:3.7197 train_time:72915ms step_avg:165.72ms
step:451/1530 train_loss:3.5827 train_time:73083ms step_avg:165.72ms
step:452/1530 train_loss:3.7139 train_time:73249ms step_avg:165.72ms
step:453/1530 train_loss:3.6730 train_time:73417ms step_avg:165.73ms
step:454/1530 train_loss:3.6467 train_time:73587ms step_avg:165.74ms
step:455/1530 train_loss:3.8465 train_time:73756ms step_avg:165.74ms
step:456/1530 train_loss:3.7251 train_time:73926ms step_avg:165.75ms
step:457/1530 train_loss:3.7846 train_time:74096ms step_avg:165.76ms
step:458/1530 train_loss:3.8297 train_time:74266ms step_avg:165.77ms
step:459/1530 train_loss:3.6332 train_time:74436ms step_avg:165.78ms
step:460/1530 train_loss:3.7952 train_time:74607ms step_avg:165.79ms
step:461/1530 train_loss:3.7002 train_time:74777ms step_avg:165.80ms
step:462/1530 train_loss:3.7378 train_time:74948ms step_avg:165.81ms
step:463/1530 train_loss:3.7773 train_time:75117ms step_avg:165.82ms
step:464/1530 train_loss:3.7181 train_time:75288ms step_avg:165.83ms
step:465/1530 train_loss:3.7174 train_time:75456ms step_avg:165.84ms
step:466/1530 train_loss:3.8056 train_time:75626ms step_avg:165.85ms
step:467/1530 train_loss:3.8304 train_time:75796ms step_avg:165.85ms
step:468/1530 train_loss:3.7971 train_time:75966ms step_avg:165.86ms
step:469/1530 train_loss:3.6961 train_time:76136ms step_avg:165.87ms
step:470/1530 train_loss:3.7682 train_time:76308ms step_avg:165.89ms
step:471/1530 train_loss:3.8151 train_time:76478ms step_avg:165.90ms
step:472/1530 train_loss:3.7918 train_time:76648ms step_avg:165.91ms
step:473/1530 train_loss:3.7191 train_time:76817ms step_avg:165.91ms
step:474/1530 train_loss:3.5965 train_time:76987ms step_avg:165.92ms
step:475/1530 train_loss:4.0286 train_time:77156ms step_avg:165.93ms
step:476/1530 train_loss:3.7588 train_time:77327ms step_avg:165.94ms
step:477/1530 train_loss:3.5919 train_time:77497ms step_avg:165.95ms
step:478/1530 train_loss:3.8292 train_time:77667ms step_avg:165.96ms
step:479/1530 train_loss:3.7768 train_time:77837ms step_avg:165.96ms
step:480/1530 train_loss:3.9249 train_time:78008ms step_avg:165.98ms
step:481/1530 train_loss:3.7262 train_time:78176ms step_avg:165.98ms
step:482/1530 train_loss:3.5292 train_time:78347ms step_avg:165.99ms
step:483/1530 train_loss:3.8077 train_time:78516ms step_avg:166.00ms
step:484/1530 train_loss:3.6670 train_time:78688ms step_avg:166.01ms
step:485/1530 train_loss:3.6616 train_time:78858ms step_avg:166.02ms
step:486/1530 train_loss:3.5828 train_time:79028ms step_avg:166.03ms
step:487/1530 train_loss:3.6893 train_time:79198ms step_avg:166.03ms
step:488/1530 train_loss:3.8804 train_time:79368ms step_avg:166.04ms
step:489/1530 train_loss:3.7107 train_time:79537ms step_avg:166.05ms
step:490/1530 train_loss:3.5975 train_time:79709ms step_avg:166.06ms
step:491/1530 train_loss:3.6146 train_time:79877ms step_avg:166.06ms
step:492/1530 train_loss:3.7397 train_time:80047ms step_avg:166.07ms
step:493/1530 train_loss:3.5865 train_time:80216ms step_avg:166.08ms
step:494/1530 train_loss:3.7014 train_time:80387ms step_avg:166.09ms
step:495/1530 train_loss:3.6634 train_time:80556ms step_avg:166.10ms
step:496/1530 train_loss:3.5161 train_time:80729ms step_avg:166.11ms
step:497/1530 train_loss:3.7372 train_time:80898ms step_avg:166.11ms
step:498/1530 train_loss:3.7914 train_time:81068ms step_avg:166.12ms
step:499/1530 train_loss:3.8235 train_time:81238ms step_avg:166.13ms
step:500/1530 train_loss:3.7396 train_time:81410ms step_avg:166.14ms
step:500/1530 val_loss:3.7116 train_time:81459ms step_avg:166.24ms
step:501/1530 train_loss:3.8111 train_time:81581ms step_avg:166.15ms
step:502/1530 train_loss:3.7512 train_time:81754ms step_avg:166.17ms
step:503/1530 train_loss:3.7786 train_time:81924ms step_avg:166.18ms
step:504/1530 train_loss:3.7208 train_time:82093ms step_avg:166.18ms
step:505/1530 train_loss:3.8080 train_time:82261ms step_avg:166.18ms
step:506/1530 train_loss:3.6464 train_time:82433ms step_avg:166.20ms
step:507/1530 train_loss:3.7662 train_time:82601ms step_avg:166.20ms
step:508/1530 train_loss:3.8305 train_time:82773ms step_avg:166.21ms
step:509/1530 train_loss:3.7792 train_time:82942ms step_avg:166.22ms
step:510/1530 train_loss:3.5867 train_time:83112ms step_avg:166.22ms
step:511/1530 train_loss:3.7766 train_time:83282ms step_avg:166.23ms
step:512/1530 train_loss:3.7229 train_time:83456ms step_avg:166.25ms
step:513/1530 train_loss:3.6669 train_time:83624ms step_avg:166.25ms
step:514/1530 train_loss:3.8965 train_time:83793ms step_avg:166.26ms
step:515/1530 train_loss:3.7320 train_time:83961ms step_avg:166.26ms
step:516/1530 train_loss:4.0841 train_time:84130ms step_avg:166.27ms
step:517/1530 train_loss:3.6925 train_time:84299ms step_avg:166.27ms
step:518/1530 train_loss:3.7717 train_time:84467ms step_avg:166.27ms
step:519/1530 train_loss:3.6575 train_time:84637ms step_avg:166.28ms
step:520/1530 train_loss:3.6852 train_time:84806ms step_avg:166.29ms
step:521/1530 train_loss:3.6666 train_time:84975ms step_avg:166.29ms
step:522/1530 train_loss:3.6676 train_time:85145ms step_avg:166.30ms
step:523/1530 train_loss:4.2908 train_time:85315ms step_avg:166.31ms
step:524/1530 train_loss:3.7377 train_time:85484ms step_avg:166.31ms
step:525/1530 train_loss:3.6825 train_time:85655ms step_avg:166.32ms
step:526/1530 train_loss:3.7042 train_time:85822ms step_avg:166.32ms
step:527/1530 train_loss:3.6582 train_time:85992ms step_avg:166.33ms
step:528/1530 train_loss:3.6288 train_time:86161ms step_avg:166.33ms
step:529/1530 train_loss:3.8525 train_time:86331ms step_avg:166.34ms
step:530/1530 train_loss:3.6489 train_time:86501ms step_avg:166.35ms
step:531/1530 train_loss:3.9241 train_time:86672ms step_avg:166.36ms
step:532/1530 train_loss:3.7375 train_time:86840ms step_avg:166.36ms
step:533/1530 train_loss:3.6579 train_time:87008ms step_avg:166.36ms
step:534/1530 train_loss:3.6739 train_time:87177ms step_avg:166.37ms
step:535/1530 train_loss:3.6094 train_time:87347ms step_avg:166.38ms
step:536/1530 train_loss:3.7530 train_time:87518ms step_avg:166.38ms
step:537/1530 train_loss:3.7304 train_time:87688ms step_avg:166.39ms
step:538/1530 train_loss:3.6271 train_time:87858ms step_avg:166.40ms
step:539/1530 train_loss:4.1256 train_time:88029ms step_avg:166.41ms
step:540/1530 train_loss:3.6775 train_time:88198ms step_avg:166.41ms
step:541/1530 train_loss:3.7911 train_time:88366ms step_avg:166.41ms
step:542/1530 train_loss:3.5882 train_time:88536ms step_avg:166.42ms
step:543/1530 train_loss:3.5878 train_time:88704ms step_avg:166.42ms
step:544/1530 train_loss:3.6411 train_time:88873ms step_avg:166.43ms
step:545/1530 train_loss:3.5922 train_time:89042ms step_avg:166.43ms
step:546/1530 train_loss:3.6320 train_time:89213ms step_avg:166.44ms
step:547/1530 train_loss:3.6479 train_time:89381ms step_avg:166.45ms
step:548/1530 train_loss:3.6142 train_time:89552ms step_avg:166.45ms
step:549/1530 train_loss:3.7285 train_time:89720ms step_avg:166.46ms
step:550/1530 train_loss:3.6242 train_time:89890ms step_avg:166.46ms
step:551/1530 train_loss:3.6346 train_time:90058ms step_avg:166.47ms
step:552/1530 train_loss:3.9390 train_time:90229ms step_avg:166.47ms
step:553/1530 train_loss:3.7628 train_time:90397ms step_avg:166.48ms
step:554/1530 train_loss:3.7190 train_time:90566ms step_avg:166.48ms
step:555/1530 train_loss:3.6354 train_time:90735ms step_avg:166.49ms
step:556/1530 train_loss:3.6981 train_time:90903ms step_avg:166.49ms
step:557/1530 train_loss:3.3156 train_time:91072ms step_avg:166.49ms
step:558/1530 train_loss:3.6131 train_time:91241ms step_avg:166.50ms
step:559/1530 train_loss:3.6491 train_time:91410ms step_avg:166.50ms
step:560/1530 train_loss:3.6928 train_time:91579ms step_avg:166.51ms
step:561/1530 train_loss:3.6126 train_time:91748ms step_avg:166.51ms
step:562/1530 train_loss:3.5541 train_time:91917ms step_avg:166.52ms
step:563/1530 train_loss:3.7620 train_time:92086ms step_avg:166.52ms
step:564/1530 train_loss:3.5752 train_time:92257ms step_avg:166.53ms
step:565/1530 train_loss:3.6813 train_time:92425ms step_avg:166.53ms
step:566/1530 train_loss:3.6202 train_time:92724ms step_avg:166.77ms
step:567/1530 train_loss:3.6038 train_time:92903ms step_avg:166.79ms
step:568/1530 train_loss:3.6889 train_time:93074ms step_avg:166.80ms
step:569/1530 train_loss:3.6470 train_time:93397ms step_avg:167.08ms
step:570/1530 train_loss:3.6837 train_time:93566ms step_avg:167.08ms
step:571/1530 train_loss:3.7659 train_time:93737ms step_avg:167.09ms
step:572/1530 train_loss:3.7336 train_time:93909ms step_avg:167.10ms
step:573/1530 train_loss:3.7372 train_time:94080ms step_avg:167.11ms
step:574/1530 train_loss:3.7844 train_time:94255ms step_avg:167.12ms
step:575/1530 train_loss:3.7283 train_time:94425ms step_avg:167.12ms
step:576/1530 train_loss:3.7619 train_time:94596ms step_avg:167.13ms
step:577/1530 train_loss:3.6687 train_time:94767ms step_avg:167.14ms
step:578/1530 train_loss:3.6831 train_time:94940ms step_avg:167.15ms
step:579/1530 train_loss:3.6740 train_time:95111ms step_avg:167.15ms
step:580/1530 train_loss:3.5902 train_time:95281ms step_avg:167.16ms
step:581/1530 train_loss:3.6424 train_time:95455ms step_avg:167.17ms
step:582/1530 train_loss:3.8553 train_time:95625ms step_avg:167.18ms
step:583/1530 train_loss:3.6294 train_time:95797ms step_avg:167.18ms
step:584/1530 train_loss:3.5958 train_time:95967ms step_avg:167.19ms
step:585/1530 train_loss:3.7918 train_time:96138ms step_avg:167.20ms
step:586/1530 train_loss:3.5195 train_time:96311ms step_avg:167.21ms
step:587/1530 train_loss:3.6750 train_time:96482ms step_avg:167.21ms
step:588/1530 train_loss:3.6432 train_time:96654ms step_avg:167.22ms
step:589/1530 train_loss:3.9927 train_time:96824ms step_avg:167.23ms
step:590/1530 train_loss:3.7804 train_time:96996ms step_avg:167.24ms
step:591/1530 train_loss:3.5165 train_time:97167ms step_avg:167.24ms
step:592/1530 train_loss:3.5364 train_time:97341ms step_avg:167.25ms
step:593/1530 train_loss:3.5074 train_time:97515ms step_avg:167.26ms
step:594/1530 train_loss:3.5588 train_time:97687ms step_avg:167.27ms
step:595/1530 train_loss:3.9239 train_time:97861ms step_avg:167.28ms
step:596/1530 train_loss:3.6500 train_time:98035ms step_avg:167.30ms
step:597/1530 train_loss:3.5878 train_time:98204ms step_avg:167.30ms
step:598/1530 train_loss:3.6576 train_time:98376ms step_avg:167.31ms
step:599/1530 train_loss:3.4838 train_time:98546ms step_avg:167.31ms
step:600/1530 train_loss:3.5967 train_time:98717ms step_avg:167.32ms
step:601/1530 train_loss:3.6489 train_time:98891ms step_avg:167.33ms
step:602/1530 train_loss:3.6710 train_time:99062ms step_avg:167.33ms
step:603/1530 train_loss:3.7808 train_time:99235ms step_avg:167.34ms
step:604/1530 train_loss:3.6127 train_time:99405ms step_avg:167.35ms
step:605/1530 train_loss:3.6182 train_time:99579ms step_avg:167.36ms
step:606/1530 train_loss:3.5791 train_time:99752ms step_avg:167.37ms
step:607/1530 train_loss:3.8425 train_time:99924ms step_avg:167.38ms
step:608/1530 train_loss:3.6375 train_time:100096ms step_avg:167.38ms
step:609/1530 train_loss:3.6163 train_time:100266ms step_avg:167.39ms
step:610/1530 train_loss:3.6983 train_time:100438ms step_avg:167.40ms
step:611/1530 train_loss:3.6022 train_time:100607ms step_avg:167.40ms
step:612/1530 train_loss:3.5726 train_time:100778ms step_avg:167.40ms
step:613/1530 train_loss:3.7634 train_time:100950ms step_avg:167.41ms
step:614/1530 train_loss:3.7062 train_time:101121ms step_avg:167.42ms
step:615/1530 train_loss:3.6972 train_time:101292ms step_avg:167.42ms
step:616/1530 train_loss:3.6328 train_time:101462ms step_avg:167.43ms
step:617/1530 train_loss:3.5562 train_time:101635ms step_avg:167.44ms
step:618/1530 train_loss:3.6891 train_time:101806ms step_avg:167.44ms
step:619/1530 train_loss:3.5531 train_time:101977ms step_avg:167.45ms
step:620/1530 train_loss:3.5960 train_time:102148ms step_avg:167.46ms
step:621/1530 train_loss:3.9275 train_time:102321ms step_avg:167.46ms
step:622/1530 train_loss:3.5757 train_time:102494ms step_avg:167.47ms
step:623/1530 train_loss:3.6057 train_time:102665ms step_avg:167.48ms
step:624/1530 train_loss:3.6917 train_time:102838ms step_avg:167.49ms
step:625/1530 train_loss:3.7070 train_time:103007ms step_avg:167.49ms
step:625/1530 val_loss:3.6236 train_time:103057ms step_avg:167.57ms
step:626/1530 train_loss:3.7419 train_time:103179ms step_avg:167.50ms
step:627/1530 train_loss:3.7165 train_time:103353ms step_avg:167.51ms
step:628/1530 train_loss:3.7596 train_time:103522ms step_avg:167.51ms
step:629/1530 train_loss:3.5979 train_time:103695ms step_avg:167.52ms
step:630/1530 train_loss:3.7336 train_time:103866ms step_avg:167.53ms
step:631/1530 train_loss:3.7444 train_time:104036ms step_avg:167.53ms
step:632/1530 train_loss:3.6494 train_time:104207ms step_avg:167.53ms
step:633/1530 train_loss:3.5995 train_time:104378ms step_avg:167.54ms
step:634/1530 train_loss:3.7069 train_time:104549ms step_avg:167.55ms
step:635/1530 train_loss:3.9548 train_time:104719ms step_avg:167.55ms
step:636/1530 train_loss:3.5525 train_time:104892ms step_avg:167.56ms
step:637/1530 train_loss:3.3582 train_time:105062ms step_avg:167.56ms
step:638/1530 train_loss:3.5951 train_time:105232ms step_avg:167.57ms
step:639/1530 train_loss:3.6356 train_time:105402ms step_avg:167.57ms
step:640/1530 train_loss:3.5753 train_time:105574ms step_avg:167.58ms
step:641/1530 train_loss:3.5853 train_time:105743ms step_avg:167.58ms
step:642/1530 train_loss:3.6320 train_time:105913ms step_avg:167.58ms
step:643/1530 train_loss:3.5944 train_time:106085ms step_avg:167.59ms
step:644/1530 train_loss:3.5618 train_time:106255ms step_avg:167.59ms
step:645/1530 train_loss:3.7814 train_time:106426ms step_avg:167.60ms
step:646/1530 train_loss:3.6802 train_time:106597ms step_avg:167.61ms
step:647/1530 train_loss:3.6629 train_time:106768ms step_avg:167.61ms
step:648/1530 train_loss:3.7138 train_time:106939ms step_avg:167.62ms
step:649/1530 train_loss:3.7711 train_time:107111ms step_avg:167.62ms
step:650/1530 train_loss:3.6184 train_time:107282ms step_avg:167.63ms
step:651/1530 train_loss:3.7727 train_time:107454ms step_avg:167.63ms
step:652/1530 train_loss:3.5879 train_time:107624ms step_avg:167.64ms
step:653/1530 train_loss:3.6689 train_time:107794ms step_avg:167.64ms
step:654/1530 train_loss:3.4319 train_time:107964ms step_avg:167.65ms
step:655/1530 train_loss:3.5865 train_time:108134ms step_avg:167.65ms
step:656/1530 train_loss:3.5793 train_time:108304ms step_avg:167.65ms
step:657/1530 train_loss:3.4983 train_time:108475ms step_avg:167.66ms
step:658/1530 train_loss:3.6903 train_time:108645ms step_avg:167.66ms
step:659/1530 train_loss:3.5859 train_time:108816ms step_avg:167.67ms
step:660/1530 train_loss:3.6851 train_time:108986ms step_avg:167.67ms
step:661/1530 train_loss:3.7562 train_time:109158ms step_avg:167.68ms
step:662/1530 train_loss:3.6727 train_time:109327ms step_avg:167.68ms
step:663/1530 train_loss:3.5546 train_time:109498ms step_avg:167.68ms
step:664/1530 train_loss:3.6115 train_time:109670ms step_avg:167.69ms
step:665/1530 train_loss:3.4941 train_time:109840ms step_avg:167.69ms
step:666/1530 train_loss:3.7807 train_time:110010ms step_avg:167.70ms
step:667/1530 train_loss:3.6105 train_time:110182ms step_avg:167.70ms
step:668/1530 train_loss:3.6506 train_time:110353ms step_avg:167.71ms
step:669/1530 train_loss:3.4897 train_time:110522ms step_avg:167.71ms
step:670/1530 train_loss:3.6016 train_time:110695ms step_avg:167.72ms
step:671/1530 train_loss:3.5631 train_time:110865ms step_avg:167.72ms
step:672/1530 train_loss:3.5697 train_time:111035ms step_avg:167.73ms
step:673/1530 train_loss:3.8588 train_time:111206ms step_avg:167.73ms
step:674/1530 train_loss:3.6303 train_time:111376ms step_avg:167.74ms
step:675/1530 train_loss:3.7126 train_time:111548ms step_avg:167.74ms
step:676/1530 train_loss:3.4885 train_time:111719ms step_avg:167.75ms
step:677/1530 train_loss:3.5992 train_time:111891ms step_avg:167.75ms
step:678/1530 train_loss:3.5570 train_time:112061ms step_avg:167.76ms
step:679/1530 train_loss:3.6820 train_time:112233ms step_avg:167.76ms
step:680/1530 train_loss:3.5879 train_time:112403ms step_avg:167.77ms
step:681/1530 train_loss:3.6211 train_time:112576ms step_avg:167.77ms
step:682/1530 train_loss:3.6614 train_time:112754ms step_avg:167.79ms
step:683/1530 train_loss:3.7408 train_time:112927ms step_avg:167.80ms
step:684/1530 train_loss:3.6470 train_time:113098ms step_avg:167.80ms
step:685/1530 train_loss:3.6903 train_time:113273ms step_avg:167.81ms
step:686/1530 train_loss:3.6461 train_time:113445ms step_avg:167.82ms
step:687/1530 train_loss:3.6692 train_time:113617ms step_avg:167.82ms
step:688/1530 train_loss:3.2147 train_time:113795ms step_avg:167.84ms
step:689/1530 train_loss:3.4074 train_time:113970ms step_avg:167.85ms
step:690/1530 train_loss:3.5395 train_time:114144ms step_avg:167.86ms
step:691/1530 train_loss:3.4173 train_time:114315ms step_avg:167.86ms
step:692/1530 train_loss:3.6281 train_time:114487ms step_avg:167.87ms
step:693/1530 train_loss:3.6552 train_time:114659ms step_avg:167.88ms
step:694/1530 train_loss:3.5566 train_time:114832ms step_avg:167.88ms
step:695/1530 train_loss:3.5363 train_time:115002ms step_avg:167.89ms
step:696/1530 train_loss:3.8515 train_time:115179ms step_avg:167.90ms
step:697/1530 train_loss:3.5948 train_time:115352ms step_avg:167.91ms
step:698/1530 train_loss:3.6464 train_time:115523ms step_avg:167.91ms
step:699/1530 train_loss:3.7660 train_time:115698ms step_avg:167.92ms
step:700/1530 train_loss:3.5778 train_time:115871ms step_avg:167.93ms
step:701/1530 train_loss:3.5428 train_time:116041ms step_avg:167.93ms
step:702/1530 train_loss:3.5154 train_time:116216ms step_avg:167.94ms
step:703/1530 train_loss:3.5009 train_time:116389ms step_avg:167.95ms
step:704/1530 train_loss:3.5767 train_time:116562ms step_avg:167.96ms
step:705/1530 train_loss:3.5582 train_time:116738ms step_avg:167.97ms
step:706/1530 train_loss:3.5803 train_time:116915ms step_avg:167.98ms
step:707/1530 train_loss:3.6463 train_time:117090ms step_avg:167.99ms
step:708/1530 train_loss:3.6017 train_time:117261ms step_avg:168.00ms
step:709/1530 train_loss:3.5786 train_time:117435ms step_avg:168.00ms
step:710/1530 train_loss:3.5405 train_time:117607ms step_avg:168.01ms
step:711/1530 train_loss:3.5943 train_time:117780ms step_avg:168.02ms
step:712/1530 train_loss:3.6443 train_time:117955ms step_avg:168.03ms
step:713/1530 train_loss:3.6541 train_time:118129ms step_avg:168.03ms
step:714/1530 train_loss:3.5626 train_time:118301ms step_avg:168.04ms
step:715/1530 train_loss:3.5725 train_time:118475ms step_avg:168.05ms
step:716/1530 train_loss:3.5889 train_time:118646ms step_avg:168.05ms
step:717/1530 train_loss:3.7115 train_time:118820ms step_avg:168.06ms
step:718/1530 train_loss:3.6001 train_time:118993ms step_avg:168.07ms
step:719/1530 train_loss:3.6782 train_time:119163ms step_avg:168.07ms
step:720/1530 train_loss:3.8459 train_time:119337ms step_avg:168.08ms
step:721/1530 train_loss:3.4687 train_time:119509ms step_avg:168.09ms
step:722/1530 train_loss:3.7421 train_time:119682ms step_avg:168.09ms
step:723/1530 train_loss:3.7674 train_time:119853ms step_avg:168.10ms
step:724/1530 train_loss:3.5705 train_time:120026ms step_avg:168.10ms
step:725/1530 train_loss:3.6550 train_time:120199ms step_avg:168.11ms
step:726/1530 train_loss:3.5340 train_time:120373ms step_avg:168.12ms
step:727/1530 train_loss:3.5799 train_time:120549ms step_avg:168.13ms
step:728/1530 train_loss:3.7333 train_time:120721ms step_avg:168.14ms
step:729/1530 train_loss:3.6728 train_time:120895ms step_avg:168.14ms
step:730/1530 train_loss:3.6618 train_time:121069ms step_avg:168.15ms
step:731/1530 train_loss:3.5558 train_time:121241ms step_avg:168.16ms
step:732/1530 train_loss:3.5996 train_time:121413ms step_avg:168.16ms
step:733/1530 train_loss:3.8369 train_time:121587ms step_avg:168.17ms
step:734/1530 train_loss:3.5615 train_time:121761ms step_avg:168.18ms
step:735/1530 train_loss:3.6191 train_time:121933ms step_avg:168.18ms
step:736/1530 train_loss:3.7411 train_time:122105ms step_avg:168.19ms
step:737/1530 train_loss:3.6776 train_time:122278ms step_avg:168.20ms
step:738/1530 train_loss:3.6042 train_time:122450ms step_avg:168.20ms
step:739/1530 train_loss:3.5107 train_time:122622ms step_avg:168.21ms
step:740/1530 train_loss:4.1157 train_time:122800ms step_avg:168.22ms
step:741/1530 train_loss:3.4883 train_time:122972ms step_avg:168.22ms
step:742/1530 train_loss:3.5614 train_time:123143ms step_avg:168.23ms
step:743/1530 train_loss:3.5825 train_time:123316ms step_avg:168.23ms
step:744/1530 train_loss:3.6501 train_time:123488ms step_avg:168.24ms
step:745/1530 train_loss:3.5929 train_time:123662ms step_avg:168.25ms
step:746/1530 train_loss:3.6001 train_time:123833ms step_avg:168.25ms
step:747/1530 train_loss:3.6477 train_time:124006ms step_avg:168.26ms
step:748/1530 train_loss:3.5656 train_time:124184ms step_avg:168.27ms
step:749/1530 train_loss:3.5620 train_time:124356ms step_avg:168.28ms
step:750/1530 train_loss:3.5997 train_time:124526ms step_avg:168.28ms
step:750/1530 val_loss:3.5705 train_time:124575ms step_avg:168.34ms
step:751/1530 train_loss:3.5730 train_time:124698ms step_avg:168.28ms
step:752/1530 train_loss:3.6196 train_time:124869ms step_avg:168.29ms
step:753/1530 train_loss:3.6223 train_time:125042ms step_avg:168.29ms
step:754/1530 train_loss:3.5945 train_time:125215ms step_avg:168.30ms
step:755/1530 train_loss:3.6859 train_time:125516ms step_avg:168.48ms
step:756/1530 train_loss:3.4607 train_time:125702ms step_avg:168.50ms
step:757/1530 train_loss:3.7261 train_time:125874ms step_avg:168.51ms
step:758/1530 train_loss:3.6506 train_time:126046ms step_avg:168.51ms
step:759/1530 train_loss:3.5840 train_time:126369ms step_avg:168.72ms
step:760/1530 train_loss:3.7071 train_time:126539ms step_avg:168.72ms
step:761/1530 train_loss:3.4007 train_time:126711ms step_avg:168.72ms
step:762/1530 train_loss:3.5634 train_time:126884ms step_avg:168.73ms
step:763/1530 train_loss:3.6628 train_time:127056ms step_avg:168.73ms
step:764/1530 train_loss:3.3217 train_time:127230ms step_avg:168.74ms
step:765/1530 train_loss:3.7323 train_time:127403ms step_avg:168.75ms
step:766/1530 train_loss:3.5695 train_time:127575ms step_avg:168.75ms
step:767/1530 train_loss:3.5669 train_time:127748ms step_avg:168.76ms
step:768/1530 train_loss:3.5645 train_time:127922ms step_avg:168.76ms
step:769/1530 train_loss:3.5840 train_time:128095ms step_avg:168.77ms
step:770/1530 train_loss:3.6362 train_time:128267ms step_avg:168.77ms
step:771/1530 train_loss:3.8913 train_time:128440ms step_avg:168.78ms
step:772/1530 train_loss:3.4523 train_time:128612ms step_avg:168.78ms
step:773/1530 train_loss:3.6291 train_time:128784ms step_avg:168.79ms
step:774/1530 train_loss:3.6384 train_time:128956ms step_avg:168.79ms
step:775/1530 train_loss:3.6051 train_time:129128ms step_avg:168.80ms
step:776/1530 train_loss:3.4130 train_time:129304ms step_avg:168.80ms
step:777/1530 train_loss:3.3949 train_time:129478ms step_avg:168.81ms
step:778/1530 train_loss:3.4900 train_time:129649ms step_avg:168.81ms
step:779/1530 train_loss:3.5868 train_time:129825ms step_avg:168.82ms
step:780/1530 train_loss:3.5911 train_time:129996ms step_avg:168.83ms
step:781/1530 train_loss:3.6752 train_time:130168ms step_avg:168.83ms
step:782/1530 train_loss:3.5945 train_time:130341ms step_avg:168.84ms
step:783/1530 train_loss:3.5727 train_time:130512ms step_avg:168.84ms
step:784/1530 train_loss:3.6147 train_time:130686ms step_avg:168.85ms
step:785/1530 train_loss:3.5631 train_time:130857ms step_avg:168.85ms
step:786/1530 train_loss:3.4337 train_time:131030ms step_avg:168.85ms
step:787/1530 train_loss:3.7365 train_time:131204ms step_avg:168.86ms
step:788/1530 train_loss:3.5114 train_time:131376ms step_avg:168.86ms
step:789/1530 train_loss:3.5522 train_time:131548ms step_avg:168.87ms
step:790/1530 train_loss:3.6255 train_time:131722ms step_avg:168.87ms
step:791/1530 train_loss:3.7741 train_time:131898ms step_avg:168.88ms
step:792/1530 train_loss:3.7624 train_time:132070ms step_avg:168.89ms
step:793/1530 train_loss:3.4480 train_time:132241ms step_avg:168.89ms
step:794/1530 train_loss:3.5968 train_time:132413ms step_avg:168.89ms
step:795/1530 train_loss:3.6761 train_time:132588ms step_avg:168.90ms
step:796/1530 train_loss:3.7349 train_time:132766ms step_avg:168.91ms
step:797/1530 train_loss:3.5272 train_time:132939ms step_avg:168.92ms
step:798/1530 train_loss:3.6401 train_time:133114ms step_avg:168.93ms
step:799/1530 train_loss:3.5367 train_time:133292ms step_avg:168.94ms
step:800/1530 train_loss:3.5301 train_time:133465ms step_avg:168.94ms
step:801/1530 train_loss:3.6280 train_time:133638ms step_avg:168.95ms
step:802/1530 train_loss:3.4987 train_time:133816ms step_avg:168.96ms
step:803/1530 train_loss:3.4797 train_time:133990ms step_avg:168.97ms
step:804/1530 train_loss:3.6252 train_time:134165ms step_avg:168.97ms
step:805/1530 train_loss:3.5216 train_time:134340ms step_avg:168.98ms
step:806/1530 train_loss:3.5641 train_time:134513ms step_avg:168.99ms
step:807/1530 train_loss:3.6428 train_time:134688ms step_avg:168.99ms
step:808/1530 train_loss:3.5457 train_time:134863ms step_avg:169.00ms
step:809/1530 train_loss:3.4913 train_time:135036ms step_avg:169.01ms
step:810/1530 train_loss:3.5675 train_time:135208ms step_avg:169.01ms
step:811/1530 train_loss:3.5843 train_time:135383ms step_avg:169.02ms
step:812/1530 train_loss:3.6089 train_time:135556ms step_avg:169.02ms
step:813/1530 train_loss:3.6246 train_time:135729ms step_avg:169.03ms
step:814/1530 train_loss:3.5695 train_time:135904ms step_avg:169.03ms
step:815/1530 train_loss:3.5652 train_time:136076ms step_avg:169.04ms
step:816/1530 train_loss:3.6854 train_time:136250ms step_avg:169.04ms
step:817/1530 train_loss:3.7662 train_time:136423ms step_avg:169.05ms
step:818/1530 train_loss:3.5240 train_time:136596ms step_avg:169.05ms
step:819/1530 train_loss:3.7201 train_time:136769ms step_avg:169.06ms
step:820/1530 train_loss:3.5011 train_time:136947ms step_avg:169.07ms
step:821/1530 train_loss:3.5648 train_time:137119ms step_avg:169.07ms
step:822/1530 train_loss:3.7035 train_time:137294ms step_avg:169.08ms
step:823/1530 train_loss:3.5770 train_time:137468ms step_avg:169.09ms
step:824/1530 train_loss:3.5111 train_time:137640ms step_avg:169.09ms
step:825/1530 train_loss:3.6156 train_time:137814ms step_avg:169.10ms
step:826/1530 train_loss:3.4782 train_time:137989ms step_avg:169.10ms
step:827/1530 train_loss:3.7299 train_time:138162ms step_avg:169.11ms
step:828/1530 train_loss:3.6128 train_time:138334ms step_avg:169.11ms
step:829/1530 train_loss:3.6310 train_time:138510ms step_avg:169.12ms
step:830/1530 train_loss:3.5387 train_time:138685ms step_avg:169.13ms
step:831/1530 train_loss:3.6010 train_time:138858ms step_avg:169.13ms
step:832/1530 train_loss:3.5178 train_time:139033ms step_avg:169.14ms
step:833/1530 train_loss:3.6551 train_time:139210ms step_avg:169.15ms
step:834/1530 train_loss:3.4727 train_time:139384ms step_avg:169.15ms
step:835/1530 train_loss:3.4609 train_time:139556ms step_avg:169.16ms
step:836/1530 train_loss:3.7219 train_time:139730ms step_avg:169.17ms
step:837/1530 train_loss:3.4009 train_time:139906ms step_avg:169.17ms
step:838/1530 train_loss:3.5970 train_time:140080ms step_avg:169.18ms
step:839/1530 train_loss:3.4180 train_time:140255ms step_avg:169.19ms
step:840/1530 train_loss:3.4659 train_time:140428ms step_avg:169.19ms
step:841/1530 train_loss:3.5680 train_time:140601ms step_avg:169.20ms
step:842/1530 train_loss:3.5855 train_time:140776ms step_avg:169.20ms
step:843/1530 train_loss:3.5608 train_time:140949ms step_avg:169.21ms
step:844/1530 train_loss:3.4281 train_time:141122ms step_avg:169.21ms
step:845/1530 train_loss:3.6621 train_time:141295ms step_avg:169.22ms
step:846/1530 train_loss:3.5159 train_time:141469ms step_avg:169.22ms
step:847/1530 train_loss:3.4963 train_time:141644ms step_avg:169.23ms
step:848/1530 train_loss:3.6351 train_time:141818ms step_avg:169.23ms
step:849/1530 train_loss:3.4935 train_time:141991ms step_avg:169.24ms
step:850/1530 train_loss:3.4404 train_time:142165ms step_avg:169.24ms
step:851/1530 train_loss:3.7398 train_time:142339ms step_avg:169.25ms
step:852/1530 train_loss:3.4353 train_time:142511ms step_avg:169.25ms
step:853/1530 train_loss:3.5680 train_time:142684ms step_avg:169.26ms
step:854/1530 train_loss:3.6517 train_time:142860ms step_avg:169.27ms
step:855/1530 train_loss:3.5203 train_time:143033ms step_avg:169.27ms
step:856/1530 train_loss:3.5463 train_time:143209ms step_avg:169.28ms
step:857/1530 train_loss:3.6098 train_time:143384ms step_avg:169.28ms
step:858/1530 train_loss:3.4706 train_time:143559ms step_avg:169.29ms
step:859/1530 train_loss:3.5637 train_time:143732ms step_avg:169.30ms
step:860/1530 train_loss:3.5880 train_time:143906ms step_avg:169.30ms
step:861/1530 train_loss:3.6324 train_time:144083ms step_avg:169.31ms
step:862/1530 train_loss:3.6043 train_time:144259ms step_avg:169.32ms
step:863/1530 train_loss:3.5704 train_time:144436ms step_avg:169.33ms
step:864/1530 train_loss:3.3875 train_time:144610ms step_avg:169.33ms
step:865/1530 train_loss:3.6020 train_time:144782ms step_avg:169.34ms
step:866/1530 train_loss:3.8987 train_time:144956ms step_avg:169.34ms
step:867/1530 train_loss:3.4606 train_time:145129ms step_avg:169.35ms
step:868/1530 train_loss:3.6510 train_time:145300ms step_avg:169.35ms
step:869/1530 train_loss:3.6113 train_time:145473ms step_avg:169.35ms
step:870/1530 train_loss:3.4512 train_time:145648ms step_avg:169.36ms
step:871/1530 train_loss:3.3889 train_time:145825ms step_avg:169.37ms
step:872/1530 train_loss:3.6512 train_time:145999ms step_avg:169.37ms
step:873/1530 train_loss:3.4605 train_time:146171ms step_avg:169.38ms
step:874/1530 train_loss:3.2218 train_time:146350ms step_avg:169.39ms
step:875/1530 train_loss:3.6345 train_time:146524ms step_avg:169.39ms
step:875/1530 val_loss:3.5210 train_time:146574ms step_avg:169.45ms
step:876/1530 train_loss:3.4391 train_time:146698ms step_avg:169.40ms
step:877/1530 train_loss:3.6228 train_time:146876ms step_avg:169.41ms
step:878/1530 train_loss:3.4699 train_time:147052ms step_avg:169.42ms
step:879/1530 train_loss:3.6551 train_time:147225ms step_avg:169.42ms
step:880/1530 train_loss:3.3132 train_time:147396ms step_avg:169.42ms
step:881/1530 train_loss:3.4823 train_time:147569ms step_avg:169.43ms
step:882/1530 train_loss:3.6985 train_time:147741ms step_avg:169.43ms
step:883/1530 train_loss:3.8406 train_time:147915ms step_avg:169.43ms
step:884/1530 train_loss:3.5672 train_time:148090ms step_avg:169.44ms
step:885/1530 train_loss:3.5021 train_time:148263ms step_avg:169.44ms
step:886/1530 train_loss:3.5676 train_time:148438ms step_avg:169.45ms
step:887/1530 train_loss:4.0858 train_time:148614ms step_avg:169.46ms
step:888/1530 train_loss:3.8379 train_time:148794ms step_avg:169.47ms
step:889/1530 train_loss:3.5247 train_time:148968ms step_avg:169.47ms
step:890/1530 train_loss:3.5336 train_time:149139ms step_avg:169.48ms
step:891/1530 train_loss:3.3584 train_time:149314ms step_avg:169.48ms
step:892/1530 train_loss:3.7159 train_time:149488ms step_avg:169.49ms
step:893/1530 train_loss:3.4269 train_time:149661ms step_avg:169.49ms
step:894/1530 train_loss:3.6399 train_time:149838ms step_avg:169.50ms
step:895/1530 train_loss:3.6814 train_time:150012ms step_avg:169.50ms
step:896/1530 train_loss:3.5008 train_time:150185ms step_avg:169.51ms
step:897/1530 train_loss:3.5436 train_time:150359ms step_avg:169.51ms
step:898/1530 train_loss:3.5932 train_time:150536ms step_avg:169.52ms
step:899/1530 train_loss:3.4818 train_time:150708ms step_avg:169.53ms
step:900/1530 train_loss:3.4268 train_time:150881ms step_avg:169.53ms
step:901/1530 train_loss:3.6186 train_time:151054ms step_avg:169.53ms
step:902/1530 train_loss:3.6328 train_time:151227ms step_avg:169.54ms
step:903/1530 train_loss:3.5436 train_time:151402ms step_avg:169.54ms
step:904/1530 train_loss:3.4914 train_time:151576ms step_avg:169.55ms
step:905/1530 train_loss:3.5019 train_time:151748ms step_avg:169.55ms
step:906/1530 train_loss:3.7108 train_time:151923ms step_avg:169.56ms
step:907/1530 train_loss:3.5212 train_time:152097ms step_avg:169.56ms
step:908/1530 train_loss:3.5695 train_time:152270ms step_avg:169.57ms
step:909/1530 train_loss:3.4548 train_time:152447ms step_avg:169.57ms
step:910/1530 train_loss:3.5307 train_time:152628ms step_avg:169.59ms
step:911/1530 train_loss:3.6457 train_time:152803ms step_avg:169.59ms
step:912/1530 train_loss:3.6001 train_time:152981ms step_avg:169.60ms
step:913/1530 train_loss:3.4604 train_time:153161ms step_avg:169.61ms
step:914/1530 train_loss:3.7493 train_time:153340ms step_avg:169.62ms
step:915/1530 train_loss:3.5411 train_time:153520ms step_avg:169.64ms
step:916/1530 train_loss:3.6208 train_time:153696ms step_avg:169.64ms
step:917/1530 train_loss:3.6002 train_time:153872ms step_avg:169.65ms
step:918/1530 train_loss:4.8307 train_time:154054ms step_avg:169.66ms
step:919/1530 train_loss:3.5009 train_time:154234ms step_avg:169.67ms
step:920/1530 train_loss:3.5935 train_time:154410ms step_avg:169.68ms
step:921/1530 train_loss:3.5516 train_time:154586ms step_avg:169.69ms
step:922/1530 train_loss:3.5894 train_time:154763ms step_avg:169.70ms
step:923/1530 train_loss:3.6176 train_time:154940ms step_avg:169.70ms
step:924/1530 train_loss:3.6789 train_time:155117ms step_avg:169.71ms
step:925/1530 train_loss:3.6480 train_time:155292ms step_avg:169.72ms
step:926/1530 train_loss:3.5602 train_time:155465ms step_avg:169.72ms
step:927/1530 train_loss:3.5592 train_time:155640ms step_avg:169.73ms
step:928/1530 train_loss:3.7826 train_time:155818ms step_avg:169.74ms
step:929/1530 train_loss:3.6118 train_time:155992ms step_avg:169.74ms
step:930/1530 train_loss:3.4032 train_time:156167ms step_avg:169.75ms
step:931/1530 train_loss:3.4936 train_time:156341ms step_avg:169.75ms
step:932/1530 train_loss:3.6499 train_time:156518ms step_avg:169.76ms
step:933/1530 train_loss:3.3614 train_time:156694ms step_avg:169.77ms
step:934/1530 train_loss:3.5810 train_time:156873ms step_avg:169.78ms
step:935/1530 train_loss:3.4410 train_time:157051ms step_avg:169.78ms
step:936/1530 train_loss:3.5242 train_time:157227ms step_avg:169.79ms
step:937/1530 train_loss:3.6230 train_time:157404ms step_avg:169.80ms
step:938/1530 train_loss:3.5434 train_time:157578ms step_avg:169.80ms
step:939/1530 train_loss:3.6764 train_time:157759ms step_avg:169.82ms
step:940/1530 train_loss:3.4838 train_time:157935ms step_avg:169.82ms
step:941/1530 train_loss:3.5525 train_time:158109ms step_avg:169.83ms
step:942/1530 train_loss:3.3600 train_time:158285ms step_avg:169.83ms
step:943/1530 train_loss:3.7086 train_time:158465ms step_avg:169.84ms
step:944/1530 train_loss:3.4000 train_time:158774ms step_avg:169.99ms
step:945/1530 train_loss:3.4282 train_time:158957ms step_avg:170.01ms
step:946/1530 train_loss:5.0916 train_time:159139ms step_avg:170.02ms
step:947/1530 train_loss:3.6009 train_time:159314ms step_avg:170.03ms
step:948/1530 train_loss:3.4886 train_time:159491ms step_avg:170.03ms
step:949/1530 train_loss:3.3748 train_time:159816ms step_avg:170.20ms
step:950/1530 train_loss:3.4420 train_time:159989ms step_avg:170.20ms
step:951/1530 train_loss:3.4097 train_time:160166ms step_avg:170.21ms
step:952/1530 train_loss:3.4800 train_time:160341ms step_avg:170.21ms
step:953/1530 train_loss:3.5675 train_time:160518ms step_avg:170.22ms
step:954/1530 train_loss:3.4511 train_time:160697ms step_avg:170.23ms
step:955/1530 train_loss:3.4740 train_time:160872ms step_avg:170.24ms
step:956/1530 train_loss:3.4430 train_time:161047ms step_avg:170.24ms
step:957/1530 train_loss:3.4952 train_time:161227ms step_avg:170.25ms
step:958/1530 train_loss:3.5105 train_time:161405ms step_avg:170.26ms
step:959/1530 train_loss:3.5143 train_time:161581ms step_avg:170.26ms
step:960/1530 train_loss:3.4039 train_time:161758ms step_avg:170.27ms
step:961/1530 train_loss:3.6452 train_time:161934ms step_avg:170.28ms
step:962/1530 train_loss:3.5931 train_time:162108ms step_avg:170.28ms
step:963/1530 train_loss:3.7891 train_time:162283ms step_avg:170.29ms
step:964/1530 train_loss:3.4310 train_time:162461ms step_avg:170.29ms
step:965/1530 train_loss:3.4771 train_time:162635ms step_avg:170.30ms
step:966/1530 train_loss:3.7064 train_time:162810ms step_avg:170.30ms
step:967/1530 train_loss:3.5314 train_time:162984ms step_avg:170.31ms
step:968/1530 train_loss:3.5169 train_time:163159ms step_avg:170.31ms
step:969/1530 train_loss:3.5850 train_time:163335ms step_avg:170.32ms
step:970/1530 train_loss:3.3811 train_time:163507ms step_avg:170.32ms
step:971/1530 train_loss:3.5341 train_time:163681ms step_avg:170.32ms
step:972/1530 train_loss:3.4733 train_time:163855ms step_avg:170.33ms
step:973/1530 train_loss:3.5395 train_time:164030ms step_avg:170.33ms
step:974/1530 train_loss:3.5889 train_time:164207ms step_avg:170.34ms
step:975/1530 train_loss:3.4657 train_time:164382ms step_avg:170.34ms
step:976/1530 train_loss:3.6698 train_time:164556ms step_avg:170.35ms
step:977/1530 train_loss:3.5763 train_time:164731ms step_avg:170.35ms
step:978/1530 train_loss:3.3679 train_time:164906ms step_avg:170.36ms
step:979/1530 train_loss:3.6293 train_time:165083ms step_avg:170.36ms
step:980/1530 train_loss:3.4177 train_time:165260ms step_avg:170.37ms
step:981/1530 train_loss:3.5748 train_time:165439ms step_avg:170.38ms
step:982/1530 train_loss:3.5435 train_time:165613ms step_avg:170.38ms
step:983/1530 train_loss:3.5236 train_time:165790ms step_avg:170.39ms
step:984/1530 train_loss:3.5002 train_time:165963ms step_avg:170.39ms
step:985/1530 train_loss:3.5793 train_time:166141ms step_avg:170.40ms
step:986/1530 train_loss:3.4118 train_time:166317ms step_avg:170.41ms
step:987/1530 train_loss:3.4873 train_time:166491ms step_avg:170.41ms
step:988/1530 train_loss:3.4797 train_time:166666ms step_avg:170.41ms
step:989/1530 train_loss:3.4134 train_time:166838ms step_avg:170.42ms
step:990/1530 train_loss:3.6616 train_time:167016ms step_avg:170.42ms
step:991/1530 train_loss:3.4750 train_time:167191ms step_avg:170.43ms
step:992/1530 train_loss:3.4444 train_time:167371ms step_avg:170.44ms
step:993/1530 train_loss:3.5006 train_time:167550ms step_avg:170.45ms
step:994/1530 train_loss:3.5980 train_time:167725ms step_avg:170.45ms
step:995/1530 train_loss:3.5316 train_time:167898ms step_avg:170.45ms
step:996/1530 train_loss:3.4608 train_time:168072ms step_avg:170.46ms
step:997/1530 train_loss:3.7525 train_time:168248ms step_avg:170.46ms
step:998/1530 train_loss:3.4399 train_time:168420ms step_avg:170.47ms
step:999/1530 train_loss:3.5870 train_time:168595ms step_avg:170.47ms
step:1000/1530 train_loss:3.4404 train_time:168773ms step_avg:170.48ms
step:1000/1530 val_loss:3.4696 train_time:168826ms step_avg:170.53ms
step:1001/1530 train_loss:3.5051 train_time:168951ms step_avg:170.48ms
step:1002/1530 train_loss:3.3813 train_time:169123ms step_avg:170.49ms
step:1003/1530 train_loss:3.5598 train_time:169299ms step_avg:170.49ms
step:1004/1530 train_loss:3.6041 train_time:169476ms step_avg:170.50ms
step:1005/1530 train_loss:3.3920 train_time:169650ms step_avg:170.50ms
step:1006/1530 train_loss:3.4677 train_time:169824ms step_avg:170.51ms
step:1007/1530 train_loss:3.4399 train_time:169999ms step_avg:170.51ms
step:1008/1530 train_loss:3.5589 train_time:170176ms step_avg:170.52ms
step:1009/1530 train_loss:3.6646 train_time:170355ms step_avg:170.53ms
step:1010/1530 train_loss:3.5615 train_time:170527ms step_avg:170.53ms
step:1011/1530 train_loss:3.5392 train_time:170702ms step_avg:170.53ms
step:1012/1530 train_loss:3.3916 train_time:170877ms step_avg:170.54ms
step:1013/1530 train_loss:3.5382 train_time:171055ms step_avg:170.54ms
step:1014/1530 train_loss:3.6272 train_time:171230ms step_avg:170.55ms
step:1015/1530 train_loss:3.3289 train_time:171407ms step_avg:170.55ms
step:1016/1530 train_loss:3.4147 train_time:171580ms step_avg:170.56ms
step:1017/1530 train_loss:3.3948 train_time:171758ms step_avg:170.56ms
step:1018/1530 train_loss:3.3944 train_time:171934ms step_avg:170.57ms
step:1019/1530 train_loss:3.5223 train_time:172111ms step_avg:170.58ms
step:1020/1530 train_loss:3.3816 train_time:172287ms step_avg:170.58ms
step:1021/1530 train_loss:3.3564 train_time:172464ms step_avg:170.59ms
step:1022/1530 train_loss:3.4779 train_time:172641ms step_avg:170.59ms
step:1023/1530 train_loss:3.5081 train_time:172817ms step_avg:170.60ms
step:1024/1530 train_loss:3.4815 train_time:172995ms step_avg:170.61ms
step:1025/1530 train_loss:3.4774 train_time:173174ms step_avg:170.61ms
step:1026/1530 train_loss:3.6198 train_time:173349ms step_avg:170.62ms
step:1027/1530 train_loss:3.3204 train_time:173525ms step_avg:170.62ms
step:1028/1530 train_loss:3.4022 train_time:173704ms step_avg:170.63ms
step:1029/1530 train_loss:3.3121 train_time:173885ms step_avg:170.64ms
step:1030/1530 train_loss:3.5390 train_time:174062ms step_avg:170.65ms
step:1031/1530 train_loss:3.5116 train_time:174240ms step_avg:170.66ms
step:1032/1530 train_loss:3.6938 train_time:174422ms step_avg:170.67ms
step:1033/1530 train_loss:3.4913 train_time:174597ms step_avg:170.67ms
step:1034/1530 train_loss:3.3956 train_time:174776ms step_avg:170.68ms
step:1035/1530 train_loss:3.4474 train_time:174955ms step_avg:170.69ms
step:1036/1530 train_loss:3.4814 train_time:175133ms step_avg:170.69ms
step:1037/1530 train_loss:3.7893 train_time:175308ms step_avg:170.70ms
step:1038/1530 train_loss:3.6168 train_time:175485ms step_avg:170.71ms
step:1039/1530 train_loss:3.5142 train_time:175667ms step_avg:170.72ms
step:1040/1530 train_loss:3.4150 train_time:175844ms step_avg:170.72ms
step:1041/1530 train_loss:3.4889 train_time:176022ms step_avg:170.73ms
step:1042/1530 train_loss:3.5227 train_time:176196ms step_avg:170.73ms
step:1043/1530 train_loss:3.4464 train_time:176372ms step_avg:170.74ms
step:1044/1530 train_loss:3.4587 train_time:176547ms step_avg:170.74ms
step:1045/1530 train_loss:3.5182 train_time:176726ms step_avg:170.75ms
step:1046/1530 train_loss:3.4246 train_time:176900ms step_avg:170.75ms
step:1047/1530 train_loss:3.6281 train_time:177077ms step_avg:170.76ms
step:1048/1530 train_loss:3.4963 train_time:177255ms step_avg:170.77ms
step:1049/1530 train_loss:3.4016 train_time:177432ms step_avg:170.77ms
step:1050/1530 train_loss:3.3949 train_time:177609ms step_avg:170.78ms
step:1051/1530 train_loss:3.4939 train_time:177786ms step_avg:170.78ms
step:1052/1530 train_loss:3.3631 train_time:177964ms step_avg:170.79ms
step:1053/1530 train_loss:3.6918 train_time:178142ms step_avg:170.80ms
step:1054/1530 train_loss:3.5324 train_time:178322ms step_avg:170.81ms
step:1055/1530 train_loss:3.3833 train_time:178496ms step_avg:170.81ms
step:1056/1530 train_loss:3.4978 train_time:178671ms step_avg:170.81ms
step:1057/1530 train_loss:3.5789 train_time:178849ms step_avg:170.82ms
step:1058/1530 train_loss:3.3049 train_time:179027ms step_avg:170.83ms
step:1059/1530 train_loss:3.3767 train_time:179206ms step_avg:170.84ms
step:1060/1530 train_loss:3.4383 train_time:179382ms step_avg:170.84ms
step:1061/1530 train_loss:3.4199 train_time:179558ms step_avg:170.85ms
step:1062/1530 train_loss:3.3846 train_time:179734ms step_avg:170.85ms
step:1063/1530 train_loss:3.4574 train_time:179908ms step_avg:170.85ms
step:1064/1530 train_loss:3.3852 train_time:180081ms step_avg:170.86ms
step:1065/1530 train_loss:3.3579 train_time:180259ms step_avg:170.86ms
step:1066/1530 train_loss:3.4119 train_time:180436ms step_avg:170.87ms
step:1067/1530 train_loss:3.2918 train_time:180615ms step_avg:170.88ms
step:1068/1530 train_loss:3.4315 train_time:180792ms step_avg:170.88ms
step:1069/1530 train_loss:3.2915 train_time:180971ms step_avg:170.89ms
step:1070/1530 train_loss:3.5661 train_time:181147ms step_avg:170.89ms
step:1071/1530 train_loss:3.5152 train_time:181327ms step_avg:170.90ms
step:1072/1530 train_loss:3.4388 train_time:181501ms step_avg:170.90ms
step:1073/1530 train_loss:3.5244 train_time:181676ms step_avg:170.91ms
step:1074/1530 train_loss:3.4317 train_time:181854ms step_avg:170.92ms
step:1075/1530 train_loss:3.3966 train_time:182030ms step_avg:170.92ms
step:1076/1530 train_loss:3.8008 train_time:182206ms step_avg:170.92ms
step:1077/1530 train_loss:3.4380 train_time:182380ms step_avg:170.93ms
step:1078/1530 train_loss:3.0933 train_time:182565ms step_avg:170.94ms
step:1079/1530 train_loss:3.5350 train_time:182742ms step_avg:170.95ms
step:1080/1530 train_loss:3.4249 train_time:182919ms step_avg:170.95ms
step:1081/1530 train_loss:3.5042 train_time:183093ms step_avg:170.96ms
step:1082/1530 train_loss:3.5900 train_time:183269ms step_avg:170.96ms
step:1083/1530 train_loss:3.4966 train_time:183444ms step_avg:170.96ms
step:1084/1530 train_loss:3.4614 train_time:183620ms step_avg:170.97ms
step:1085/1530 train_loss:3.4319 train_time:183795ms step_avg:170.97ms
step:1086/1530 train_loss:3.6278 train_time:183971ms step_avg:170.98ms
step:1087/1530 train_loss:3.5036 train_time:184147ms step_avg:170.98ms
step:1088/1530 train_loss:3.3705 train_time:184323ms step_avg:170.99ms
step:1089/1530 train_loss:3.3746 train_time:184502ms step_avg:170.99ms
step:1090/1530 train_loss:3.4831 train_time:184680ms step_avg:171.00ms
step:1091/1530 train_loss:3.2860 train_time:184858ms step_avg:171.01ms
step:1092/1530 train_loss:3.4869 train_time:185036ms step_avg:171.01ms
step:1093/1530 train_loss:3.5996 train_time:185215ms step_avg:171.02ms
step:1094/1530 train_loss:3.4493 train_time:185391ms step_avg:171.02ms
step:1095/1530 train_loss:3.4154 train_time:185565ms step_avg:171.03ms
step:1096/1530 train_loss:3.4253 train_time:185743ms step_avg:171.03ms
step:1097/1530 train_loss:3.4886 train_time:185921ms step_avg:171.04ms
step:1098/1530 train_loss:3.5668 train_time:186098ms step_avg:171.05ms
step:1099/1530 train_loss:3.5319 train_time:186275ms step_avg:171.05ms
step:1100/1530 train_loss:3.4265 train_time:186454ms step_avg:171.06ms
step:1101/1530 train_loss:3.2963 train_time:186632ms step_avg:171.07ms
step:1102/1530 train_loss:3.3067 train_time:186810ms step_avg:171.07ms
step:1103/1530 train_loss:3.4395 train_time:186992ms step_avg:171.08ms
step:1104/1530 train_loss:3.3212 train_time:187168ms step_avg:171.09ms
step:1105/1530 train_loss:4.0577 train_time:187347ms step_avg:171.09ms
step:1106/1530 train_loss:3.2276 train_time:187521ms step_avg:171.10ms
step:1107/1530 train_loss:3.5704 train_time:187696ms step_avg:171.10ms
step:1108/1530 train_loss:3.3478 train_time:187870ms step_avg:171.10ms
step:1109/1530 train_loss:3.5006 train_time:188045ms step_avg:171.11ms
step:1110/1530 train_loss:3.4293 train_time:188219ms step_avg:171.11ms
step:1111/1530 train_loss:3.4859 train_time:188393ms step_avg:171.11ms
step:1112/1530 train_loss:3.5577 train_time:188571ms step_avg:171.12ms
step:1113/1530 train_loss:3.4355 train_time:188754ms step_avg:171.13ms
step:1114/1530 train_loss:3.3681 train_time:188934ms step_avg:171.14ms
step:1115/1530 train_loss:3.2389 train_time:189113ms step_avg:171.14ms
step:1116/1530 train_loss:3.4341 train_time:189285ms step_avg:171.14ms
step:1117/1530 train_loss:3.5931 train_time:189466ms step_avg:171.15ms
step:1118/1530 train_loss:3.6287 train_time:189643ms step_avg:171.16ms
step:1119/1530 train_loss:3.4814 train_time:189819ms step_avg:171.16ms
step:1120/1530 train_loss:3.4943 train_time:189996ms step_avg:171.17ms
step:1121/1530 train_loss:3.3930 train_time:190173ms step_avg:171.17ms
step:1122/1530 train_loss:3.4569 train_time:190349ms step_avg:171.18ms
step:1123/1530 train_loss:3.5783 train_time:190523ms step_avg:171.18ms
step:1124/1530 train_loss:3.3424 train_time:190699ms step_avg:171.18ms
step:1125/1530 train_loss:3.2244 train_time:190876ms step_avg:171.19ms
step:1125/1530 val_loss:3.4087 train_time:190927ms step_avg:171.23ms
step:1126/1530 train_loss:3.4771 train_time:191053ms step_avg:171.19ms
step:1127/1530 train_loss:3.6723 train_time:191231ms step_avg:171.20ms
step:1128/1530 train_loss:3.2316 train_time:191409ms step_avg:171.21ms
step:1129/1530 train_loss:3.5594 train_time:191588ms step_avg:171.21ms
step:1130/1530 train_loss:3.3779 train_time:191769ms step_avg:171.22ms
step:1131/1530 train_loss:3.4017 train_time:191951ms step_avg:171.23ms
step:1132/1530 train_loss:3.3677 train_time:192124ms step_avg:171.23ms
step:1133/1530 train_loss:3.4917 train_time:192430ms step_avg:171.35ms
step:1134/1530 train_loss:3.4438 train_time:192618ms step_avg:171.37ms
step:1135/1530 train_loss:3.5236 train_time:192793ms step_avg:171.37ms
step:1136/1530 train_loss:3.5634 train_time:192971ms step_avg:171.38ms
step:1137/1530 train_loss:3.4581 train_time:193147ms step_avg:171.38ms
step:1138/1530 train_loss:3.3555 train_time:193328ms step_avg:171.39ms
step:1139/1530 train_loss:3.6578 train_time:193655ms step_avg:171.53ms
step:1140/1530 train_loss:3.4539 train_time:193833ms step_avg:171.53ms
step:1141/1530 train_loss:3.5996 train_time:194013ms step_avg:171.54ms
step:1142/1530 train_loss:3.4457 train_time:194189ms step_avg:171.54ms
step:1143/1530 train_loss:3.3652 train_time:194367ms step_avg:171.55ms
step:1144/1530 train_loss:3.4439 train_time:194546ms step_avg:171.56ms
step:1145/1530 train_loss:3.5889 train_time:194721ms step_avg:171.56ms
step:1146/1530 train_loss:3.5582 train_time:194901ms step_avg:171.57ms
step:1147/1530 train_loss:3.4892 train_time:195077ms step_avg:171.57ms
step:1148/1530 train_loss:3.4953 train_time:195256ms step_avg:171.58ms
step:1149/1530 train_loss:3.3261 train_time:195437ms step_avg:171.59ms
step:1150/1530 train_loss:3.3761 train_time:195612ms step_avg:171.59ms
step:1151/1530 train_loss:3.3223 train_time:195791ms step_avg:171.60ms
step:1152/1530 train_loss:3.3979 train_time:195973ms step_avg:171.60ms
step:1153/1530 train_loss:3.4311 train_time:196153ms step_avg:171.61ms
step:1154/1530 train_loss:3.5175 train_time:196330ms step_avg:171.62ms
step:1155/1530 train_loss:3.3127 train_time:196511ms step_avg:171.63ms
step:1156/1530 train_loss:3.5366 train_time:196696ms step_avg:171.64ms
step:1157/1530 train_loss:3.4946 train_time:196872ms step_avg:171.64ms
step:1158/1530 train_loss:3.2504 train_time:197050ms step_avg:171.65ms
step:1159/1530 train_loss:3.3486 train_time:197227ms step_avg:171.65ms
step:1160/1530 train_loss:3.3392 train_time:197400ms step_avg:171.65ms
step:1161/1530 train_loss:3.0850 train_time:197580ms step_avg:171.66ms
step:1162/1530 train_loss:3.4219 train_time:197759ms step_avg:171.67ms
step:1163/1530 train_loss:3.3927 train_time:197939ms step_avg:171.67ms
step:1164/1530 train_loss:3.2921 train_time:198116ms step_avg:171.68ms
step:1165/1530 train_loss:3.2503 train_time:198293ms step_avg:171.68ms
step:1166/1530 train_loss:3.3915 train_time:198472ms step_avg:171.69ms
step:1167/1530 train_loss:3.4159 train_time:198649ms step_avg:171.69ms
step:1168/1530 train_loss:3.7250 train_time:198824ms step_avg:171.70ms
step:1169/1530 train_loss:3.3769 train_time:199002ms step_avg:171.70ms
step:1170/1530 train_loss:3.3900 train_time:199179ms step_avg:171.71ms
step:1171/1530 train_loss:3.3040 train_time:199356ms step_avg:171.71ms
step:1172/1530 train_loss:3.4244 train_time:199530ms step_avg:171.71ms
step:1173/1530 train_loss:3.5388 train_time:199710ms step_avg:171.72ms
step:1174/1530 train_loss:3.3815 train_time:199894ms step_avg:171.73ms
step:1175/1530 train_loss:3.3647 train_time:200075ms step_avg:171.74ms
step:1176/1530 train_loss:3.4271 train_time:200258ms step_avg:171.75ms
step:1177/1530 train_loss:3.4489 train_time:200440ms step_avg:171.76ms
step:1178/1530 train_loss:3.4972 train_time:200615ms step_avg:171.76ms
step:1179/1530 train_loss:3.4011 train_time:200790ms step_avg:171.76ms
step:1180/1530 train_loss:3.3587 train_time:200976ms step_avg:171.77ms
step:1181/1530 train_loss:3.3411 train_time:201153ms step_avg:171.78ms
step:1182/1530 train_loss:3.3744 train_time:201331ms step_avg:171.78ms
step:1183/1530 train_loss:3.3342 train_time:201506ms step_avg:171.79ms
step:1184/1530 train_loss:3.5073 train_time:201684ms step_avg:171.79ms
step:1185/1530 train_loss:3.5421 train_time:201868ms step_avg:171.80ms
step:1186/1530 train_loss:3.3681 train_time:202047ms step_avg:171.81ms
step:1187/1530 train_loss:3.4160 train_time:202233ms step_avg:171.82ms
step:1188/1530 train_loss:3.4417 train_time:202409ms step_avg:171.82ms
step:1189/1530 train_loss:3.2770 train_time:202589ms step_avg:171.83ms
step:1190/1530 train_loss:3.4470 train_time:202768ms step_avg:171.84ms
step:1191/1530 train_loss:3.5849 train_time:202949ms step_avg:171.85ms
step:1192/1530 train_loss:3.3924 train_time:203125ms step_avg:171.85ms
step:1193/1530 train_loss:3.2746 train_time:203300ms step_avg:171.85ms
step:1194/1530 train_loss:3.5569 train_time:203478ms step_avg:171.86ms
step:1195/1530 train_loss:3.3708 train_time:203659ms step_avg:171.86ms
step:1196/1530 train_loss:3.3849 train_time:203845ms step_avg:171.88ms
step:1197/1530 train_loss:3.2934 train_time:204025ms step_avg:171.88ms
step:1198/1530 train_loss:3.3020 train_time:204209ms step_avg:171.89ms
step:1199/1530 train_loss:3.3402 train_time:204390ms step_avg:171.90ms
step:1200/1530 train_loss:3.4483 train_time:204568ms step_avg:171.91ms
step:1201/1530 train_loss:3.4807 train_time:204747ms step_avg:171.91ms
step:1202/1530 train_loss:3.6039 train_time:204935ms step_avg:171.93ms
step:1203/1530 train_loss:3.4041 train_time:205115ms step_avg:171.93ms
step:1204/1530 train_loss:3.3078 train_time:205294ms step_avg:171.94ms
step:1205/1530 train_loss:3.4390 train_time:205472ms step_avg:171.94ms
step:1206/1530 train_loss:3.4769 train_time:205649ms step_avg:171.95ms
step:1207/1530 train_loss:3.5187 train_time:205827ms step_avg:171.95ms
step:1208/1530 train_loss:3.3948 train_time:206003ms step_avg:171.96ms
step:1209/1530 train_loss:3.2434 train_time:206185ms step_avg:171.96ms
step:1210/1530 train_loss:3.3029 train_time:206363ms step_avg:171.97ms
step:1211/1530 train_loss:3.3941 train_time:206540ms step_avg:171.97ms
step:1212/1530 train_loss:3.3956 train_time:206717ms step_avg:171.98ms
step:1213/1530 train_loss:3.4114 train_time:206896ms step_avg:171.98ms
step:1214/1530 train_loss:3.2541 train_time:207077ms step_avg:171.99ms
step:1215/1530 train_loss:3.3987 train_time:207253ms step_avg:171.99ms
step:1216/1530 train_loss:3.3308 train_time:207430ms step_avg:172.00ms
step:1217/1530 train_loss:3.3263 train_time:207607ms step_avg:172.00ms
step:1218/1530 train_loss:3.4077 train_time:207787ms step_avg:172.01ms
step:1219/1530 train_loss:3.2547 train_time:207972ms step_avg:172.02ms
step:1220/1530 train_loss:3.4697 train_time:208148ms step_avg:172.02ms
step:1221/1530 train_loss:3.5094 train_time:208325ms step_avg:172.03ms
step:1222/1530 train_loss:3.4314 train_time:208500ms step_avg:172.03ms
step:1223/1530 train_loss:3.3002 train_time:208679ms step_avg:172.04ms
step:1224/1530 train_loss:3.2488 train_time:208862ms step_avg:172.04ms
step:1225/1530 train_loss:3.3657 train_time:209040ms step_avg:172.05ms
step:1226/1530 train_loss:3.3322 train_time:209221ms step_avg:172.06ms
step:1227/1530 train_loss:3.2775 train_time:209402ms step_avg:172.06ms
step:1228/1530 train_loss:3.4479 train_time:209579ms step_avg:172.07ms
step:1229/1530 train_loss:3.3678 train_time:209757ms step_avg:172.07ms
step:1230/1530 train_loss:3.4026 train_time:209940ms step_avg:172.08ms
step:1231/1530 train_loss:3.5803 train_time:210119ms step_avg:172.09ms
step:1232/1530 train_loss:3.4996 train_time:210300ms step_avg:172.09ms
step:1233/1530 train_loss:3.4252 train_time:210477ms step_avg:172.10ms
step:1234/1530 train_loss:3.5862 train_time:210654ms step_avg:172.10ms
step:1235/1530 train_loss:3.3252 train_time:210835ms step_avg:172.11ms
step:1236/1530 train_loss:3.2915 train_time:211010ms step_avg:172.11ms
step:1237/1530 train_loss:3.2716 train_time:211188ms step_avg:172.12ms
step:1238/1530 train_loss:3.2777 train_time:211370ms step_avg:172.13ms
step:1239/1530 train_loss:3.3329 train_time:211549ms step_avg:172.13ms
step:1240/1530 train_loss:3.3840 train_time:211726ms step_avg:172.13ms
step:1241/1530 train_loss:3.4282 train_time:211904ms step_avg:172.14ms
step:1242/1530 train_loss:3.2976 train_time:212084ms step_avg:172.15ms
step:1243/1530 train_loss:3.4088 train_time:212264ms step_avg:172.15ms
step:1244/1530 train_loss:3.4035 train_time:212439ms step_avg:172.15ms
step:1245/1530 train_loss:3.4081 train_time:212614ms step_avg:172.16ms
step:1246/1530 train_loss:3.2379 train_time:212792ms step_avg:172.16ms
step:1247/1530 train_loss:3.3721 train_time:212968ms step_avg:172.17ms
step:1248/1530 train_loss:3.4266 train_time:213145ms step_avg:172.17ms
step:1249/1530 train_loss:3.4254 train_time:213325ms step_avg:172.18ms
step:1250/1530 train_loss:3.3080 train_time:213505ms step_avg:172.18ms
step:1250/1530 val_loss:3.3555 train_time:213561ms step_avg:172.23ms
step:1251/1530 train_loss:3.4960 train_time:213691ms step_avg:172.19ms
step:1252/1530 train_loss:3.3597 train_time:213867ms step_avg:172.20ms
step:1253/1530 train_loss:3.3093 train_time:214046ms step_avg:172.20ms
step:1254/1530 train_loss:3.4193 train_time:214228ms step_avg:172.21ms
step:1255/1530 train_loss:3.5207 train_time:214417ms step_avg:172.22ms
step:1256/1530 train_loss:3.3123 train_time:214600ms step_avg:172.23ms
step:1257/1530 train_loss:3.3763 train_time:214778ms step_avg:172.24ms
step:1258/1530 train_loss:3.3654 train_time:214963ms step_avg:172.25ms
step:1259/1530 train_loss:3.3334 train_time:215142ms step_avg:172.25ms
step:1260/1530 train_loss:3.2076 train_time:215319ms step_avg:172.26ms
step:1261/1530 train_loss:3.3039 train_time:215498ms step_avg:172.26ms
step:1262/1530 train_loss:3.3261 train_time:215682ms step_avg:172.27ms
step:1263/1530 train_loss:3.2384 train_time:215864ms step_avg:172.28ms
step:1264/1530 train_loss:3.4445 train_time:216041ms step_avg:172.28ms
step:1265/1530 train_loss:3.4251 train_time:216218ms step_avg:172.29ms
step:1266/1530 train_loss:3.4448 train_time:216397ms step_avg:172.29ms
step:1267/1530 train_loss:3.3730 train_time:216576ms step_avg:172.30ms
step:1268/1530 train_loss:3.4095 train_time:216757ms step_avg:172.30ms
step:1269/1530 train_loss:3.2565 train_time:216943ms step_avg:172.31ms
step:1270/1530 train_loss:3.1099 train_time:217121ms step_avg:172.32ms
step:1271/1530 train_loss:3.4073 train_time:217300ms step_avg:172.32ms
step:1272/1530 train_loss:3.3538 train_time:217474ms step_avg:172.33ms
step:1273/1530 train_loss:3.3776 train_time:217656ms step_avg:172.33ms
step:1274/1530 train_loss:3.3597 train_time:217838ms step_avg:172.34ms
step:1275/1530 train_loss:3.4369 train_time:218013ms step_avg:172.34ms
step:1276/1530 train_loss:3.4719 train_time:218186ms step_avg:172.34ms
step:1277/1530 train_loss:3.4109 train_time:218366ms step_avg:172.35ms
step:1278/1530 train_loss:3.4073 train_time:218541ms step_avg:172.35ms
step:1279/1530 train_loss:3.2656 train_time:218724ms step_avg:172.36ms
step:1280/1530 train_loss:3.3664 train_time:218908ms step_avg:172.37ms
step:1281/1530 train_loss:3.4243 train_time:219086ms step_avg:172.37ms
step:1282/1530 train_loss:3.4706 train_time:219262ms step_avg:172.38ms
step:1283/1530 train_loss:3.3342 train_time:219443ms step_avg:172.38ms
step:1284/1530 train_loss:3.3720 train_time:219621ms step_avg:172.39ms
step:1285/1530 train_loss:3.3647 train_time:219800ms step_avg:172.39ms
step:1286/1530 train_loss:3.3365 train_time:219977ms step_avg:172.40ms
step:1287/1530 train_loss:3.4902 train_time:220156ms step_avg:172.40ms
step:1288/1530 train_loss:3.2940 train_time:220337ms step_avg:172.41ms
step:1289/1530 train_loss:3.3831 train_time:220523ms step_avg:172.42ms
step:1290/1530 train_loss:3.4620 train_time:220707ms step_avg:172.43ms
step:1291/1530 train_loss:3.3852 train_time:220887ms step_avg:172.43ms
step:1292/1530 train_loss:3.4815 train_time:221068ms step_avg:172.44ms
step:1293/1530 train_loss:3.5158 train_time:221249ms step_avg:172.45ms
step:1294/1530 train_loss:3.4595 train_time:221429ms step_avg:172.45ms
step:1295/1530 train_loss:3.2830 train_time:221608ms step_avg:172.46ms
step:1296/1530 train_loss:3.3780 train_time:221790ms step_avg:172.46ms
step:1297/1530 train_loss:3.2809 train_time:221969ms step_avg:172.47ms
step:1298/1530 train_loss:3.2773 train_time:222152ms step_avg:172.48ms
step:1299/1530 train_loss:3.3954 train_time:222330ms step_avg:172.48ms
step:1300/1530 train_loss:3.4023 train_time:222505ms step_avg:172.48ms
step:1301/1530 train_loss:3.4025 train_time:222683ms step_avg:172.49ms
step:1302/1530 train_loss:3.5745 train_time:222866ms step_avg:172.50ms
step:1303/1530 train_loss:3.3037 train_time:223049ms step_avg:172.51ms
step:1304/1530 train_loss:3.5134 train_time:223230ms step_avg:172.51ms
step:1305/1530 train_loss:3.2599 train_time:223406ms step_avg:172.51ms
step:1306/1530 train_loss:3.4522 train_time:223586ms step_avg:172.52ms
step:1307/1530 train_loss:3.4549 train_time:223761ms step_avg:172.52ms
step:1308/1530 train_loss:3.2891 train_time:223940ms step_avg:172.53ms
step:1309/1530 train_loss:3.3080 train_time:224121ms step_avg:172.53ms
step:1310/1530 train_loss:3.2880 train_time:224299ms step_avg:172.54ms
step:1311/1530 train_loss:3.2941 train_time:224477ms step_avg:172.54ms
step:1312/1530 train_loss:3.3761 train_time:224659ms step_avg:172.55ms
step:1313/1530 train_loss:3.3448 train_time:224835ms step_avg:172.55ms
step:1314/1530 train_loss:3.0422 train_time:225017ms step_avg:172.56ms
step:1315/1530 train_loss:3.2785 train_time:225195ms step_avg:172.56ms
step:1316/1530 train_loss:3.3987 train_time:225370ms step_avg:172.56ms
step:1317/1530 train_loss:3.4235 train_time:225548ms step_avg:172.57ms
step:1318/1530 train_loss:3.3027 train_time:225734ms step_avg:172.58ms
step:1319/1530 train_loss:3.4299 train_time:225913ms step_avg:172.58ms
step:1320/1530 train_loss:3.4633 train_time:226095ms step_avg:172.59ms
step:1321/1530 train_loss:3.3665 train_time:226274ms step_avg:172.60ms
step:1322/1530 train_loss:3.3272 train_time:226584ms step_avg:172.70ms
step:1323/1530 train_loss:3.3241 train_time:226775ms step_avg:172.72ms
step:1324/1530 train_loss:3.4400 train_time:226958ms step_avg:172.72ms
step:1325/1530 train_loss:3.4933 train_time:227143ms step_avg:172.73ms
step:1326/1530 train_loss:3.2134 train_time:227323ms step_avg:172.74ms
step:1327/1530 train_loss:3.1700 train_time:227500ms step_avg:172.74ms
step:1328/1530 train_loss:3.4976 train_time:227677ms step_avg:172.74ms
step:1329/1530 train_loss:3.3010 train_time:228017ms step_avg:172.87ms
step:1330/1530 train_loss:3.4283 train_time:228198ms step_avg:172.88ms
step:1331/1530 train_loss:3.3325 train_time:228375ms step_avg:172.88ms
step:1332/1530 train_loss:3.7420 train_time:228555ms step_avg:172.89ms
step:1333/1530 train_loss:3.4774 train_time:228735ms step_avg:172.89ms
step:1334/1530 train_loss:3.3687 train_time:228913ms step_avg:172.89ms
step:1335/1530 train_loss:3.2930 train_time:229091ms step_avg:172.90ms
step:1336/1530 train_loss:3.2966 train_time:229276ms step_avg:172.91ms
step:1337/1530 train_loss:3.5519 train_time:229456ms step_avg:172.91ms
step:1338/1530 train_loss:3.5240 train_time:229634ms step_avg:172.92ms
step:1339/1530 train_loss:3.3432 train_time:229812ms step_avg:172.92ms
step:1340/1530 train_loss:3.2895 train_time:229991ms step_avg:172.93ms
step:1341/1530 train_loss:3.5905 train_time:230167ms step_avg:172.93ms
step:1342/1530 train_loss:3.3608 train_time:230348ms step_avg:172.93ms
step:1343/1530 train_loss:3.3662 train_time:230525ms step_avg:172.94ms
step:1344/1530 train_loss:3.4185 train_time:230705ms step_avg:172.94ms
step:1345/1530 train_loss:3.3865 train_time:230889ms step_avg:172.95ms
step:1346/1530 train_loss:3.3020 train_time:231067ms step_avg:172.95ms
step:1347/1530 train_loss:3.2781 train_time:231245ms step_avg:172.96ms
step:1348/1530 train_loss:3.3501 train_time:231423ms step_avg:172.96ms
step:1349/1530 train_loss:3.2788 train_time:231598ms step_avg:172.96ms
step:1350/1530 train_loss:3.3935 train_time:231778ms step_avg:172.97ms
step:1351/1530 train_loss:3.2439 train_time:231956ms step_avg:172.97ms
step:1352/1530 train_loss:3.3125 train_time:232134ms step_avg:172.98ms
step:1353/1530 train_loss:3.4042 train_time:232314ms step_avg:172.98ms
step:1354/1530 train_loss:3.2618 train_time:232491ms step_avg:172.98ms
step:1355/1530 train_loss:3.1906 train_time:232669ms step_avg:172.99ms
step:1356/1530 train_loss:3.5115 train_time:232850ms step_avg:172.99ms
step:1357/1530 train_loss:3.4266 train_time:233030ms step_avg:173.00ms
step:1358/1530 train_loss:3.1896 train_time:233208ms step_avg:173.00ms
step:1359/1530 train_loss:3.4449 train_time:233387ms step_avg:173.01ms
step:1360/1530 train_loss:3.3536 train_time:233568ms step_avg:173.01ms
step:1361/1530 train_loss:3.1236 train_time:233754ms step_avg:173.02ms
step:1362/1530 train_loss:3.3926 train_time:233936ms step_avg:173.03ms
step:1363/1530 train_loss:3.2852 train_time:234123ms step_avg:173.04ms
step:1364/1530 train_loss:3.3049 train_time:234299ms step_avg:173.04ms
step:1365/1530 train_loss:3.3127 train_time:234478ms step_avg:173.05ms
step:1366/1530 train_loss:3.4220 train_time:234660ms step_avg:173.05ms
step:1367/1530 train_loss:3.4015 train_time:234838ms step_avg:173.06ms
step:1368/1530 train_loss:3.3474 train_time:235017ms step_avg:173.06ms
step:1369/1530 train_loss:3.2808 train_time:235206ms step_avg:173.07ms
step:1370/1530 train_loss:3.6096 train_time:235387ms step_avg:173.08ms
step:1371/1530 train_loss:3.3154 train_time:235569ms step_avg:173.08ms
step:1372/1530 train_loss:3.3735 train_time:235751ms step_avg:173.09ms
step:1373/1530 train_loss:3.3754 train_time:235931ms step_avg:173.10ms
step:1374/1530 train_loss:3.1520 train_time:236111ms step_avg:173.10ms
step:1375/1530 train_loss:3.5370 train_time:236290ms step_avg:173.11ms
step:1375/1530 val_loss:3.3132 train_time:236341ms step_avg:173.14ms
step:1376/1530 train_loss:3.3497 train_time:236470ms step_avg:173.11ms
step:1377/1530 train_loss:3.4816 train_time:236648ms step_avg:173.12ms
step:1378/1530 train_loss:3.4751 train_time:236825ms step_avg:173.12ms
step:1379/1530 train_loss:3.1180 train_time:237006ms step_avg:173.12ms
step:1380/1530 train_loss:3.3131 train_time:237186ms step_avg:173.13ms
step:1381/1530 train_loss:3.7040 train_time:237372ms step_avg:173.14ms
step:1382/1530 train_loss:3.2100 train_time:237553ms step_avg:173.14ms
step:1383/1530 train_loss:3.3953 train_time:237733ms step_avg:173.15ms
step:1384/1530 train_loss:3.4810 train_time:237916ms step_avg:173.16ms
step:1385/1530 train_loss:3.4117 train_time:238091ms step_avg:173.16ms
step:1386/1530 train_loss:3.3382 train_time:238270ms step_avg:173.16ms
step:1387/1530 train_loss:3.1985 train_time:238449ms step_avg:173.17ms
step:1388/1530 train_loss:3.3457 train_time:238627ms step_avg:173.17ms
step:1389/1530 train_loss:3.3167 train_time:238809ms step_avg:173.18ms
step:1390/1530 train_loss:3.5721 train_time:238986ms step_avg:173.18ms
step:1391/1530 train_loss:3.2926 train_time:239163ms step_avg:173.18ms
step:1392/1530 train_loss:3.2895 train_time:239343ms step_avg:173.19ms
step:1393/1530 train_loss:3.2396 train_time:239523ms step_avg:173.19ms
step:1394/1530 train_loss:3.4975 train_time:239700ms step_avg:173.19ms
step:1395/1530 train_loss:3.3893 train_time:239879ms step_avg:173.20ms
step:1396/1530 train_loss:3.4042 train_time:240058ms step_avg:173.20ms
step:1397/1530 train_loss:3.3085 train_time:240235ms step_avg:173.20ms
step:1398/1530 train_loss:3.2591 train_time:240410ms step_avg:173.21ms
step:1399/1530 train_loss:3.3218 train_time:240589ms step_avg:173.21ms
step:1400/1530 train_loss:3.3196 train_time:240773ms step_avg:173.22ms
step:1401/1530 train_loss:3.3509 train_time:240949ms step_avg:173.22ms
step:1402/1530 train_loss:3.3029 train_time:241127ms step_avg:173.22ms
step:1403/1530 train_loss:3.4984 train_time:241312ms step_avg:173.23ms
step:1404/1530 train_loss:3.2819 train_time:241490ms step_avg:173.24ms
step:1405/1530 train_loss:3.3129 train_time:241671ms step_avg:173.24ms
step:1406/1530 train_loss:3.3181 train_time:241855ms step_avg:173.25ms
step:1407/1530 train_loss:3.1755 train_time:242032ms step_avg:173.25ms
step:1408/1530 train_loss:3.3141 train_time:242212ms step_avg:173.26ms
step:1409/1530 train_loss:3.3024 train_time:242398ms step_avg:173.27ms
step:1410/1530 train_loss:3.2867 train_time:242576ms step_avg:173.27ms
step:1411/1530 train_loss:3.3690 train_time:242753ms step_avg:173.27ms
step:1412/1530 train_loss:3.3388 train_time:242928ms step_avg:173.27ms
step:1413/1530 train_loss:3.3657 train_time:243107ms step_avg:173.28ms
step:1414/1530 train_loss:3.3293 train_time:243287ms step_avg:173.28ms
step:1415/1530 train_loss:3.4134 train_time:243473ms step_avg:173.29ms
step:1416/1530 train_loss:3.2297 train_time:243662ms step_avg:173.30ms
step:1417/1530 train_loss:3.2801 train_time:243845ms step_avg:173.31ms
step:1418/1530 train_loss:3.3937 train_time:244025ms step_avg:173.31ms
step:1419/1530 train_loss:3.3502 train_time:244208ms step_avg:173.32ms
step:1420/1530 train_loss:3.3684 train_time:244388ms step_avg:173.32ms
step:1421/1530 train_loss:3.3730 train_time:244567ms step_avg:173.33ms
step:1422/1530 train_loss:3.3329 train_time:244744ms step_avg:173.33ms
step:1423/1530 train_loss:3.3188 train_time:244924ms step_avg:173.34ms
step:1424/1530 train_loss:3.3375 train_time:245109ms step_avg:173.34ms
step:1425/1530 train_loss:3.1925 train_time:245299ms step_avg:173.36ms
step:1426/1530 train_loss:3.3249 train_time:245477ms step_avg:173.36ms
step:1427/1530 train_loss:3.2876 train_time:245662ms step_avg:173.37ms
step:1428/1530 train_loss:3.3779 train_time:245840ms step_avg:173.37ms
step:1429/1530 train_loss:3.3540 train_time:246019ms step_avg:173.38ms
step:1430/1530 train_loss:3.2617 train_time:246201ms step_avg:173.38ms
step:1431/1530 train_loss:3.3231 train_time:246382ms step_avg:173.39ms
step:1432/1530 train_loss:3.3407 train_time:246565ms step_avg:173.39ms
step:1433/1530 train_loss:3.1298 train_time:246749ms step_avg:173.40ms
step:1434/1530 train_loss:3.2888 train_time:246935ms step_avg:173.41ms
step:1435/1530 train_loss:3.1205 train_time:247116ms step_avg:173.41ms
step:1436/1530 train_loss:3.2341 train_time:247294ms step_avg:173.42ms
step:1437/1530 train_loss:3.4100 train_time:247471ms step_avg:173.42ms
step:1438/1530 train_loss:3.3856 train_time:247647ms step_avg:173.42ms
step:1439/1530 train_loss:3.3139 train_time:247826ms step_avg:173.43ms
step:1440/1530 train_loss:3.1919 train_time:248003ms step_avg:173.43ms
step:1441/1530 train_loss:3.3410 train_time:248182ms step_avg:173.43ms
step:1442/1530 train_loss:3.3945 train_time:248366ms step_avg:173.44ms
step:1443/1530 train_loss:3.4879 train_time:248554ms step_avg:173.45ms
step:1444/1530 train_loss:3.4507 train_time:248731ms step_avg:173.45ms
step:1445/1530 train_loss:3.3386 train_time:248908ms step_avg:173.46ms
step:1446/1530 train_loss:3.2001 train_time:249088ms step_avg:173.46ms
step:1447/1530 train_loss:3.2940 train_time:249269ms step_avg:173.47ms
step:1448/1530 train_loss:3.3003 train_time:249448ms step_avg:173.47ms
step:1449/1530 train_loss:3.3986 train_time:249626ms step_avg:173.47ms
step:1450/1530 train_loss:3.3959 train_time:249807ms step_avg:173.48ms
step:1451/1530 train_loss:3.2073 train_time:249985ms step_avg:173.48ms
step:1452/1530 train_loss:3.3274 train_time:250166ms step_avg:173.49ms
step:1453/1530 train_loss:3.2651 train_time:250340ms step_avg:173.49ms
step:1454/1530 train_loss:3.2884 train_time:250519ms step_avg:173.49ms
step:1455/1530 train_loss:3.3316 train_time:250702ms step_avg:173.50ms
step:1456/1530 train_loss:3.2838 train_time:250879ms step_avg:173.50ms
step:1457/1530 train_loss:3.1614 train_time:251056ms step_avg:173.50ms
step:1458/1530 train_loss:3.4252 train_time:251233ms step_avg:173.50ms
step:1459/1530 train_loss:3.2778 train_time:251415ms step_avg:173.51ms
step:1460/1530 train_loss:3.3180 train_time:251595ms step_avg:173.51ms
step:1461/1530 train_loss:3.4295 train_time:251775ms step_avg:173.52ms
step:1462/1530 train_loss:3.2618 train_time:251952ms step_avg:173.52ms
step:1463/1530 train_loss:3.4681 train_time:252135ms step_avg:173.53ms
step:1464/1530 train_loss:3.3624 train_time:252313ms step_avg:173.53ms
step:1465/1530 train_loss:3.3568 train_time:252493ms step_avg:173.53ms
step:1466/1530 train_loss:3.2883 train_time:252670ms step_avg:173.54ms
step:1467/1530 train_loss:3.3973 train_time:252852ms step_avg:173.54ms
step:1468/1530 train_loss:3.2856 train_time:253029ms step_avg:173.55ms
step:1469/1530 train_loss:3.2752 train_time:253208ms step_avg:173.55ms
step:1470/1530 train_loss:3.3302 train_time:253391ms step_avg:173.56ms
step:1471/1530 train_loss:3.2583 train_time:253577ms step_avg:173.56ms
step:1472/1530 train_loss:3.2513 train_time:253761ms step_avg:173.57ms
step:1473/1530 train_loss:3.4431 train_time:253939ms step_avg:173.57ms
step:1474/1530 train_loss:3.3147 train_time:254123ms step_avg:173.58ms
step:1475/1530 train_loss:3.1485 train_time:254308ms step_avg:173.59ms
step:1476/1530 train_loss:3.2699 train_time:254487ms step_avg:173.59ms
step:1477/1530 train_loss:3.2362 train_time:254675ms step_avg:173.60ms
step:1478/1530 train_loss:3.3107 train_time:254860ms step_avg:173.61ms
step:1479/1530 train_loss:3.3996 train_time:255043ms step_avg:173.62ms
step:1480/1530 train_loss:3.2761 train_time:255221ms step_avg:173.62ms
step:1481/1530 train_loss:3.4524 train_time:255403ms step_avg:173.63ms
step:1482/1530 train_loss:3.3702 train_time:255590ms step_avg:173.63ms
step:1483/1530 train_loss:3.2795 train_time:255783ms step_avg:173.65ms
step:1484/1530 train_loss:3.2654 train_time:255972ms step_avg:173.66ms
step:1485/1530 train_loss:3.2809 train_time:256153ms step_avg:173.66ms
step:1486/1530 train_loss:3.2260 train_time:256340ms step_avg:173.67ms
step:1487/1530 train_loss:3.3434 train_time:256522ms step_avg:173.68ms
step:1488/1530 train_loss:3.2414 train_time:256705ms step_avg:173.68ms
step:1489/1530 train_loss:3.3174 train_time:256886ms step_avg:173.69ms
step:1490/1530 train_loss:3.2529 train_time:257067ms step_avg:173.69ms
step:1491/1530 train_loss:3.1618 train_time:257248ms step_avg:173.70ms
step:1492/1530 train_loss:3.2728 train_time:257428ms step_avg:173.70ms
step:1493/1530 train_loss:3.4387 train_time:257606ms step_avg:173.71ms
step:1494/1530 train_loss:3.2972 train_time:257786ms step_avg:173.71ms
step:1495/1530 train_loss:3.0337 train_time:257970ms step_avg:173.72ms
step:1496/1530 train_loss:3.3613 train_time:258154ms step_avg:173.72ms
step:1497/1530 train_loss:3.3124 train_time:258339ms step_avg:173.73ms
step:1498/1530 train_loss:3.3496 train_time:258524ms step_avg:173.74ms
step:1499/1530 train_loss:3.3149 train_time:258712ms step_avg:173.75ms
step:1500/1530 train_loss:3.3025 train_time:258905ms step_avg:173.76ms
step:1500/1530 val_loss:3.2817 train_time:258960ms step_avg:173.80ms
step:1501/1530 train_loss:3.0891 train_time:259095ms step_avg:173.77ms
step:1502/1530 train_loss:3.3642 train_time:259286ms step_avg:173.78ms
step:1503/1530 train_loss:3.2463 train_time:259465ms step_avg:173.79ms
step:1504/1530 train_loss:3.2514 train_time:259646ms step_avg:173.79ms
step:1505/1530 train_loss:3.2179 train_time:259823ms step_avg:173.79ms
step:1506/1530 train_loss:3.2820 train_time:260005ms step_avg:173.80ms
step:1507/1530 train_loss:3.1833 train_time:260202ms step_avg:173.82ms
step:1508/1530 train_loss:3.4889 train_time:260387ms step_avg:173.82ms
step:1509/1530 train_loss:3.2846 train_time:260564ms step_avg:173.83ms
step:1510/1530 train_loss:3.2763 train_time:260744ms step_avg:173.83ms
step:1511/1530 train_loss:3.4170 train_time:261052ms step_avg:173.92ms
step:1512/1530 train_loss:3.4194 train_time:261241ms step_avg:173.93ms
step:1513/1530 train_loss:3.2724 train_time:261428ms step_avg:173.94ms
step:1514/1530 train_loss:3.0873 train_time:261611ms step_avg:173.94ms
step:1515/1530 train_loss:3.2490 train_time:261791ms step_avg:173.95ms
step:1516/1530 train_loss:3.2623 train_time:261976ms step_avg:173.95ms
step:1517/1530 train_loss:3.3038 train_time:262157ms step_avg:173.96ms
step:1518/1530 train_loss:3.2106 train_time:262340ms step_avg:173.97ms
step:1519/1530 train_loss:3.5121 train_time:262673ms step_avg:174.07ms
step:1520/1530 train_loss:3.1303 train_time:262855ms step_avg:174.08ms
step:1521/1530 train_loss:3.2103 train_time:263033ms step_avg:174.08ms
step:1522/1530 train_loss:3.3627 train_time:263218ms step_avg:174.09ms
step:1523/1530 train_loss:3.2277 train_time:263395ms step_avg:174.09ms
step:1524/1530 train_loss:3.3491 train_time:263575ms step_avg:174.09ms
step:1525/1530 train_loss:3.3422 train_time:263763ms step_avg:174.10ms
step:1526/1530 train_loss:3.2818 train_time:263954ms step_avg:174.11ms
step:1527/1530 train_loss:3.2937 train_time:264135ms step_avg:174.12ms
step:1528/1530 train_loss:3.4117 train_time:264316ms step_avg:174.12ms
step:1529/1530 train_loss:3.4121 train_time:264494ms step_avg:174.12ms
step:1530/1530 train_loss:3.2403 train_time:264672ms step_avg:174.13ms
step:1530/1530 val_loss:3.2793 train_time:264727ms step_avg:174.16ms