385d2312-0cf9-48c3-af3f-35c12c12a38d.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 01:29:13 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   39C    P0              76W / 700W |      3MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   1  NVIDIA H100 80GB HBM3          On  | 00000000:3B:00.0 Off |                    0 |
| N/A   31C    P0             115W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   2  NVIDIA H100 80GB HBM3          On  | 00000000:4C:00.0 Off |                    0 |
| N/A   31C    P0             118W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   3  NVIDIA H100 80GB HBM3          On  | 00000000:5D:00.0 Off |                    0 |
| N/A   39C    P0             119W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   4  NVIDIA H100 80GB HBM3          On  | 00000000:9B:00.0 Off |                    0 |
| N/A   40C    P0             124W / 700W |    529MiB / 81559MiB |      0%      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   40C    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:32049ms step_avg:nanms
step:2/1530 train_loss:10.0766 train_time:32160ms step_avg:nanms
step:3/1530 train_loss:8.3879 train_time:32319ms step_avg:nanms
step:4/1530 train_loss:7.5514 train_time:32480ms step_avg:nanms
step:5/1530 train_loss:7.4792 train_time:32640ms step_avg:nanms
step:6/1530 train_loss:6.9785 train_time:32801ms step_avg:nanms
step:7/1530 train_loss:7.2038 train_time:32960ms step_avg:nanms
step:8/1530 train_loss:6.7355 train_time:33122ms step_avg:nanms
step:9/1530 train_loss:6.6138 train_time:33283ms step_avg:nanms
step:10/1530 train_loss:6.5174 train_time:33443ms step_avg:nanms
step:11/1530 train_loss:6.4147 train_time:114ms step_avg:nanms
step:12/1530 train_loss:6.3597 train_time:274ms step_avg:nanms
step:13/1530 train_loss:6.2692 train_time:434ms step_avg:144.80ms
step:14/1530 train_loss:6.2541 train_time:594ms step_avg:148.61ms
step:15/1530 train_loss:6.1734 train_time:754ms step_avg:150.89ms
step:16/1530 train_loss:6.1208 train_time:914ms step_avg:152.40ms
step:17/1530 train_loss:6.1645 train_time:1075ms step_avg:153.55ms
step:18/1530 train_loss:5.9837 train_time:1234ms step_avg:154.30ms
step:19/1530 train_loss:5.9729 train_time:1395ms step_avg:154.96ms
step:20/1530 train_loss:5.7055 train_time:1554ms step_avg:155.45ms
step:21/1530 train_loss:5.9536 train_time:1715ms step_avg:155.90ms
step:22/1530 train_loss:6.1746 train_time:1875ms step_avg:156.27ms
step:23/1530 train_loss:5.8526 train_time:2036ms step_avg:156.59ms
step:24/1530 train_loss:6.0153 train_time:2196ms step_avg:156.85ms
step:25/1530 train_loss:5.6742 train_time:2355ms step_avg:157.00ms
step:26/1530 train_loss:5.5956 train_time:2515ms step_avg:157.19ms
step:27/1530 train_loss:5.7718 train_time:2676ms step_avg:157.41ms
step:28/1530 train_loss:5.4122 train_time:2835ms step_avg:157.52ms
step:29/1530 train_loss:5.6699 train_time:2996ms step_avg:157.67ms
step:30/1530 train_loss:5.4832 train_time:3155ms step_avg:157.75ms
step:31/1530 train_loss:5.4433 train_time:3315ms step_avg:157.88ms
step:32/1530 train_loss:5.3056 train_time:3475ms step_avg:157.94ms
step:33/1530 train_loss:5.5735 train_time:3635ms step_avg:158.02ms
step:34/1530 train_loss:5.4909 train_time:3795ms step_avg:158.12ms
step:35/1530 train_loss:5.6217 train_time:3954ms step_avg:158.17ms
step:36/1530 train_loss:5.5430 train_time:4114ms step_avg:158.24ms
step:37/1530 train_loss:5.4593 train_time:4275ms step_avg:158.32ms
step:38/1530 train_loss:5.3352 train_time:4435ms step_avg:158.38ms
step:39/1530 train_loss:5.3491 train_time:4595ms step_avg:158.43ms
step:40/1530 train_loss:5.2653 train_time:4754ms step_avg:158.47ms
step:41/1530 train_loss:5.2242 train_time:4914ms step_avg:158.51ms
step:42/1530 train_loss:5.1802 train_time:5075ms step_avg:158.58ms
step:43/1530 train_loss:5.2754 train_time:5234ms step_avg:158.59ms
step:44/1530 train_loss:5.2398 train_time:5395ms step_avg:158.67ms
step:45/1530 train_loss:5.3888 train_time:5554ms step_avg:158.70ms
step:46/1530 train_loss:5.1736 train_time:5714ms step_avg:158.73ms
step:47/1530 train_loss:5.0503 train_time:5875ms step_avg:158.80ms
step:48/1530 train_loss:5.2002 train_time:6036ms step_avg:158.84ms
step:49/1530 train_loss:5.1374 train_time:6196ms step_avg:158.87ms
step:50/1530 train_loss:5.2522 train_time:6355ms step_avg:158.89ms
step:51/1530 train_loss:5.1427 train_time:6515ms step_avg:158.90ms
step:52/1530 train_loss:5.0396 train_time:6675ms step_avg:158.94ms
step:53/1530 train_loss:5.1746 train_time:6835ms step_avg:158.96ms
step:54/1530 train_loss:5.0111 train_time:6996ms step_avg:159.00ms
step:55/1530 train_loss:5.4062 train_time:7155ms step_avg:159.01ms
step:56/1530 train_loss:5.0336 train_time:7315ms step_avg:159.02ms
step:57/1530 train_loss:4.8878 train_time:7475ms step_avg:159.05ms
step:58/1530 train_loss:5.0415 train_time:7635ms step_avg:159.06ms
step:59/1530 train_loss:5.0225 train_time:7795ms step_avg:159.08ms
step:60/1530 train_loss:5.1434 train_time:7954ms step_avg:159.09ms
step:61/1530 train_loss:4.8463 train_time:8115ms step_avg:159.11ms
step:62/1530 train_loss:4.9889 train_time:8275ms step_avg:159.13ms
step:63/1530 train_loss:5.0081 train_time:8435ms step_avg:159.15ms
step:64/1530 train_loss:4.8897 train_time:8596ms step_avg:159.18ms
step:65/1530 train_loss:4.8017 train_time:8755ms step_avg:159.18ms
step:66/1530 train_loss:4.9138 train_time:8915ms step_avg:159.20ms
step:67/1530 train_loss:4.8269 train_time:9075ms step_avg:159.21ms
step:68/1530 train_loss:5.0798 train_time:9235ms step_avg:159.22ms
step:69/1530 train_loss:4.7338 train_time:9395ms step_avg:159.24ms
step:70/1530 train_loss:4.8304 train_time:9555ms step_avg:159.25ms
step:71/1530 train_loss:4.9506 train_time:9714ms step_avg:159.25ms
step:72/1530 train_loss:4.8748 train_time:9875ms step_avg:159.27ms
step:73/1530 train_loss:4.7697 train_time:10034ms step_avg:159.27ms
step:74/1530 train_loss:4.9003 train_time:10195ms step_avg:159.29ms
step:75/1530 train_loss:4.8667 train_time:10354ms step_avg:159.29ms
step:76/1530 train_loss:4.7999 train_time:10514ms step_avg:159.30ms
step:77/1530 train_loss:4.9133 train_time:10674ms step_avg:159.31ms
step:78/1530 train_loss:5.1067 train_time:10833ms step_avg:159.31ms
step:79/1530 train_loss:4.8159 train_time:10994ms step_avg:159.33ms
step:80/1530 train_loss:4.8492 train_time:11153ms step_avg:159.33ms
step:81/1530 train_loss:4.6301 train_time:11314ms step_avg:159.35ms
step:82/1530 train_loss:4.8048 train_time:11474ms step_avg:159.36ms
step:83/1530 train_loss:4.7637 train_time:11634ms step_avg:159.37ms
step:84/1530 train_loss:4.7538 train_time:11794ms step_avg:159.38ms
step:85/1530 train_loss:4.6187 train_time:11954ms step_avg:159.39ms
step:86/1530 train_loss:4.8284 train_time:12115ms step_avg:159.40ms
step:87/1530 train_loss:4.7432 train_time:12275ms step_avg:159.41ms
step:88/1530 train_loss:4.7297 train_time:12435ms step_avg:159.42ms
step:89/1530 train_loss:4.7038 train_time:12596ms step_avg:159.44ms
step:90/1530 train_loss:4.6253 train_time:12755ms step_avg:159.44ms
step:91/1530 train_loss:4.6209 train_time:12915ms step_avg:159.45ms
step:92/1530 train_loss:4.7771 train_time:13075ms step_avg:159.45ms
step:93/1530 train_loss:4.6067 train_time:13235ms step_avg:159.46ms
step:94/1530 train_loss:4.6593 train_time:13396ms step_avg:159.47ms
step:95/1530 train_loss:4.6921 train_time:13555ms step_avg:159.47ms
step:96/1530 train_loss:4.5678 train_time:13715ms step_avg:159.47ms
step:97/1530 train_loss:4.6205 train_time:13875ms step_avg:159.48ms
step:98/1530 train_loss:4.5836 train_time:14035ms step_avg:159.49ms
step:99/1530 train_loss:4.6571 train_time:14195ms step_avg:159.50ms
step:100/1530 train_loss:4.6659 train_time:14355ms step_avg:159.50ms
step:101/1530 train_loss:4.5085 train_time:14515ms step_avg:159.51ms
step:102/1530 train_loss:4.6956 train_time:14675ms step_avg:159.51ms
step:103/1530 train_loss:4.5793 train_time:14835ms step_avg:159.52ms
step:104/1530 train_loss:4.5305 train_time:14995ms step_avg:159.53ms
step:105/1530 train_loss:4.5440 train_time:15155ms step_avg:159.53ms
step:106/1530 train_loss:4.5875 train_time:15315ms step_avg:159.53ms
step:107/1530 train_loss:4.4964 train_time:15475ms step_avg:159.54ms
step:108/1530 train_loss:4.3544 train_time:15636ms step_avg:159.55ms
step:109/1530 train_loss:4.4752 train_time:15795ms step_avg:159.55ms
step:110/1530 train_loss:4.4697 train_time:15956ms step_avg:159.56ms
step:111/1530 train_loss:4.4217 train_time:16115ms step_avg:159.56ms
step:112/1530 train_loss:4.5815 train_time:16275ms step_avg:159.56ms
step:113/1530 train_loss:4.4849 train_time:16434ms step_avg:159.55ms
step:114/1530 train_loss:4.3570 train_time:16595ms step_avg:159.57ms
step:115/1530 train_loss:4.4921 train_time:16756ms step_avg:159.59ms
step:116/1530 train_loss:4.4644 train_time:16919ms step_avg:159.62ms
step:117/1530 train_loss:4.3616 train_time:17084ms step_avg:159.66ms
step:118/1530 train_loss:4.5899 train_time:17249ms step_avg:159.71ms
step:119/1530 train_loss:4.4482 train_time:17413ms step_avg:159.75ms
step:120/1530 train_loss:4.3223 train_time:17576ms step_avg:159.79ms
step:121/1530 train_loss:4.2922 train_time:17739ms step_avg:159.81ms
step:122/1530 train_loss:4.4401 train_time:17902ms step_avg:159.84ms
step:123/1530 train_loss:4.2692 train_time:18067ms step_avg:159.89ms
step:124/1530 train_loss:4.5733 train_time:18231ms step_avg:159.92ms
step:125/1530 train_loss:4.4439 train_time:18394ms step_avg:159.95ms
step:125/1530 val_loss:4.4102 train_time:18442ms step_avg:160.36ms
step:126/1530 train_loss:4.4245 train_time:18561ms step_avg:160.01ms
step:127/1530 train_loss:4.4265 train_time:18726ms step_avg:160.06ms
step:128/1530 train_loss:4.3756 train_time:18891ms step_avg:160.09ms
step:129/1530 train_loss:4.6875 train_time:19054ms step_avg:160.11ms
step:130/1530 train_loss:4.3524 train_time:19217ms step_avg:160.14ms
step:131/1530 train_loss:4.3831 train_time:19382ms step_avg:160.18ms
step:132/1530 train_loss:4.3359 train_time:19545ms step_avg:160.21ms
step:133/1530 train_loss:4.4405 train_time:19711ms step_avg:160.25ms
step:134/1530 train_loss:4.2636 train_time:19875ms step_avg:160.28ms
step:135/1530 train_loss:4.4497 train_time:20038ms step_avg:160.30ms
step:136/1530 train_loss:4.2115 train_time:20202ms step_avg:160.33ms
step:137/1530 train_loss:4.3634 train_time:20366ms step_avg:160.37ms
step:138/1530 train_loss:4.2748 train_time:20530ms step_avg:160.39ms
step:139/1530 train_loss:4.3748 train_time:20693ms step_avg:160.41ms
step:140/1530 train_loss:4.4702 train_time:20858ms step_avg:160.45ms
step:141/1530 train_loss:4.3077 train_time:21023ms step_avg:160.48ms
step:142/1530 train_loss:4.2950 train_time:21188ms step_avg:160.52ms
step:143/1530 train_loss:4.2474 train_time:21353ms step_avg:160.55ms
step:144/1530 train_loss:4.3498 train_time:21516ms step_avg:160.57ms
step:145/1530 train_loss:4.2928 train_time:21680ms step_avg:160.59ms
step:146/1530 train_loss:4.1631 train_time:21845ms step_avg:160.62ms
step:147/1530 train_loss:4.3283 train_time:22009ms step_avg:160.65ms
step:148/1530 train_loss:4.3609 train_time:22174ms step_avg:160.68ms
step:149/1530 train_loss:4.2926 train_time:22337ms step_avg:160.70ms
step:150/1530 train_loss:4.4321 train_time:22500ms step_avg:160.72ms
step:151/1530 train_loss:4.2803 train_time:22665ms step_avg:160.75ms
step:152/1530 train_loss:4.2895 train_time:22829ms step_avg:160.77ms
step:153/1530 train_loss:4.3739 train_time:22993ms step_avg:160.79ms
step:154/1530 train_loss:4.3655 train_time:23156ms step_avg:160.81ms
step:155/1530 train_loss:4.2668 train_time:23321ms step_avg:160.83ms
step:156/1530 train_loss:4.3551 train_time:23486ms step_avg:160.86ms
step:157/1530 train_loss:4.4121 train_time:23650ms step_avg:160.88ms
step:158/1530 train_loss:4.2425 train_time:23813ms step_avg:160.90ms
step:159/1530 train_loss:4.2988 train_time:23977ms step_avg:160.92ms
step:160/1530 train_loss:4.1345 train_time:24141ms step_avg:160.94ms
step:161/1530 train_loss:4.3538 train_time:24304ms step_avg:160.96ms
step:162/1530 train_loss:4.3591 train_time:24468ms step_avg:160.98ms
step:163/1530 train_loss:4.3353 train_time:24632ms step_avg:160.99ms
step:164/1530 train_loss:4.1799 train_time:24795ms step_avg:161.00ms
step:165/1530 train_loss:4.2811 train_time:24960ms step_avg:161.03ms
step:166/1530 train_loss:4.3336 train_time:25125ms step_avg:161.06ms
step:167/1530 train_loss:4.2020 train_time:25289ms step_avg:161.07ms
step:168/1530 train_loss:4.2987 train_time:25453ms step_avg:161.09ms
step:169/1530 train_loss:4.1721 train_time:25616ms step_avg:161.11ms
step:170/1530 train_loss:4.0415 train_time:25779ms step_avg:161.12ms
step:171/1530 train_loss:4.2175 train_time:25943ms step_avg:161.13ms
step:172/1530 train_loss:4.2202 train_time:26105ms step_avg:161.14ms
step:173/1530 train_loss:4.2685 train_time:26268ms step_avg:161.15ms
step:174/1530 train_loss:4.4191 train_time:26430ms step_avg:161.16ms
step:175/1530 train_loss:4.2482 train_time:26593ms step_avg:161.17ms
step:176/1530 train_loss:4.0979 train_time:26755ms step_avg:161.17ms
step:177/1530 train_loss:4.0759 train_time:26918ms step_avg:161.19ms
step:178/1530 train_loss:4.1880 train_time:27081ms step_avg:161.19ms
step:179/1530 train_loss:4.1258 train_time:27244ms step_avg:161.21ms
step:180/1530 train_loss:4.1144 train_time:27407ms step_avg:161.22ms
step:181/1530 train_loss:4.2912 train_time:27571ms step_avg:161.23ms
step:182/1530 train_loss:4.1480 train_time:27733ms step_avg:161.24ms
step:183/1530 train_loss:4.1168 train_time:27895ms step_avg:161.24ms
step:184/1530 train_loss:4.1312 train_time:28058ms step_avg:161.25ms
step:185/1530 train_loss:4.2027 train_time:28221ms step_avg:161.26ms
step:186/1530 train_loss:4.1701 train_time:28384ms step_avg:161.28ms
step:187/1530 train_loss:4.2349 train_time:28547ms step_avg:161.28ms
step:188/1530 train_loss:4.1704 train_time:28842ms step_avg:162.03ms
step:189/1530 train_loss:4.1156 train_time:29176ms step_avg:162.99ms
step:190/1530 train_loss:4.2088 train_time:29341ms step_avg:163.00ms
step:191/1530 train_loss:4.0866 train_time:29504ms step_avg:163.00ms
step:192/1530 train_loss:4.0404 train_time:29667ms step_avg:163.01ms
step:193/1530 train_loss:4.2535 train_time:29829ms step_avg:163.00ms
step:194/1530 train_loss:4.1861 train_time:29993ms step_avg:163.01ms
step:195/1530 train_loss:4.3595 train_time:30155ms step_avg:163.00ms
step:196/1530 train_loss:4.1803 train_time:30317ms step_avg:162.99ms
step:197/1530 train_loss:4.0483 train_time:30481ms step_avg:163.00ms
step:198/1530 train_loss:4.1801 train_time:30645ms step_avg:163.01ms
step:199/1530 train_loss:4.0396 train_time:30809ms step_avg:163.01ms
step:200/1530 train_loss:4.1253 train_time:30972ms step_avg:163.01ms
step:201/1530 train_loss:4.0086 train_time:31134ms step_avg:163.00ms
step:202/1530 train_loss:4.2546 train_time:31295ms step_avg:163.00ms
step:203/1530 train_loss:4.0619 train_time:31459ms step_avg:163.00ms
step:204/1530 train_loss:4.1928 train_time:31622ms step_avg:163.00ms
step:205/1530 train_loss:4.2541 train_time:31787ms step_avg:163.01ms
step:206/1530 train_loss:3.9447 train_time:31950ms step_avg:163.01ms
step:207/1530 train_loss:4.0788 train_time:32114ms step_avg:163.01ms
step:208/1530 train_loss:4.1005 train_time:32276ms step_avg:163.01ms
step:209/1530 train_loss:4.2331 train_time:32439ms step_avg:163.01ms
step:210/1530 train_loss:4.1706 train_time:32602ms step_avg:163.01ms
step:211/1530 train_loss:4.0611 train_time:32765ms step_avg:163.01ms
step:212/1530 train_loss:4.1271 train_time:32927ms step_avg:163.01ms
step:213/1530 train_loss:4.0453 train_time:33090ms step_avg:163.01ms
step:214/1530 train_loss:4.1187 train_time:33253ms step_avg:163.01ms
step:215/1530 train_loss:3.9530 train_time:33415ms step_avg:163.00ms
step:216/1530 train_loss:4.0102 train_time:33578ms step_avg:163.00ms
step:217/1530 train_loss:4.0215 train_time:33741ms step_avg:163.00ms
step:218/1530 train_loss:4.0878 train_time:33904ms step_avg:163.00ms
step:219/1530 train_loss:4.0736 train_time:34068ms step_avg:163.00ms
step:220/1530 train_loss:4.0858 train_time:34230ms step_avg:163.00ms
step:221/1530 train_loss:4.0938 train_time:34392ms step_avg:162.99ms
step:222/1530 train_loss:4.0005 train_time:34555ms step_avg:163.00ms
step:223/1530 train_loss:4.0003 train_time:34718ms step_avg:163.00ms
step:224/1530 train_loss:4.3042 train_time:34880ms step_avg:162.99ms
step:225/1530 train_loss:3.9248 train_time:35043ms step_avg:162.99ms
step:226/1530 train_loss:3.9993 train_time:35207ms step_avg:163.00ms
step:227/1530 train_loss:3.9746 train_time:35370ms step_avg:163.00ms
step:228/1530 train_loss:4.1469 train_time:35535ms step_avg:163.01ms
step:229/1530 train_loss:3.9321 train_time:35701ms step_avg:163.02ms
step:230/1530 train_loss:4.0409 train_time:35867ms step_avg:163.03ms
step:231/1530 train_loss:3.9116 train_time:36032ms step_avg:163.04ms
step:232/1530 train_loss:3.9702 train_time:36198ms step_avg:163.05ms
step:233/1530 train_loss:4.0902 train_time:36365ms step_avg:163.07ms
step:234/1530 train_loss:4.0324 train_time:36530ms step_avg:163.08ms
step:235/1530 train_loss:3.9093 train_time:36696ms step_avg:163.09ms
step:236/1530 train_loss:4.0859 train_time:36863ms step_avg:163.11ms
step:237/1530 train_loss:4.0804 train_time:37030ms step_avg:163.13ms
step:238/1530 train_loss:3.9431 train_time:37196ms step_avg:163.14ms
step:239/1530 train_loss:4.0733 train_time:37362ms step_avg:163.15ms
step:240/1530 train_loss:4.1140 train_time:37527ms step_avg:163.16ms
step:241/1530 train_loss:3.9646 train_time:37694ms step_avg:163.18ms
step:242/1530 train_loss:4.1454 train_time:37861ms step_avg:163.20ms
step:243/1530 train_loss:4.0111 train_time:38028ms step_avg:163.21ms
step:244/1530 train_loss:4.0807 train_time:38193ms step_avg:163.22ms
step:245/1530 train_loss:4.1415 train_time:38358ms step_avg:163.22ms
step:246/1530 train_loss:4.0562 train_time:38523ms step_avg:163.23ms
step:247/1530 train_loss:4.0085 train_time:38689ms step_avg:163.25ms
step:248/1530 train_loss:4.1126 train_time:38855ms step_avg:163.26ms
step:249/1530 train_loss:3.9322 train_time:39021ms step_avg:163.27ms
step:250/1530 train_loss:3.9855 train_time:39188ms step_avg:163.28ms
step:250/1530 val_loss:4.0119 train_time:39235ms step_avg:163.48ms
step:251/1530 train_loss:4.0810 train_time:39355ms step_avg:163.30ms
step:252/1530 train_loss:4.1773 train_time:39520ms step_avg:163.31ms
step:253/1530 train_loss:3.9333 train_time:39687ms step_avg:163.32ms
step:254/1530 train_loss:3.8914 train_time:39852ms step_avg:163.33ms
step:255/1530 train_loss:4.0816 train_time:40018ms step_avg:163.34ms
step:256/1530 train_loss:3.9920 train_time:40184ms step_avg:163.35ms
step:257/1530 train_loss:3.9910 train_time:40350ms step_avg:163.36ms
step:258/1530 train_loss:3.9876 train_time:40516ms step_avg:163.37ms
step:259/1530 train_loss:4.0380 train_time:40683ms step_avg:163.39ms
step:260/1530 train_loss:4.0640 train_time:40850ms step_avg:163.40ms
step:261/1530 train_loss:4.0331 train_time:41017ms step_avg:163.41ms
step:262/1530 train_loss:4.0063 train_time:41182ms step_avg:163.42ms
step:263/1530 train_loss:3.9036 train_time:41348ms step_avg:163.43ms
step:264/1530 train_loss:4.0001 train_time:41514ms step_avg:163.44ms
step:265/1530 train_loss:3.8656 train_time:41680ms step_avg:163.45ms
step:266/1530 train_loss:3.9256 train_time:41847ms step_avg:163.47ms
step:267/1530 train_loss:3.9353 train_time:42012ms step_avg:163.47ms
step:268/1530 train_loss:3.9722 train_time:42177ms step_avg:163.48ms
step:269/1530 train_loss:3.8556 train_time:42344ms step_avg:163.49ms
step:270/1530 train_loss:4.1016 train_time:42510ms step_avg:163.50ms
step:271/1530 train_loss:3.9705 train_time:42675ms step_avg:163.51ms
step:272/1530 train_loss:3.9417 train_time:42844ms step_avg:163.53ms
step:273/1530 train_loss:3.9660 train_time:43011ms step_avg:163.54ms
step:274/1530 train_loss:4.0509 train_time:43177ms step_avg:163.55ms
step:275/1530 train_loss:4.0623 train_time:43344ms step_avg:163.56ms
step:276/1530 train_loss:4.2354 train_time:43510ms step_avg:163.57ms
step:277/1530 train_loss:4.0423 train_time:43676ms step_avg:163.58ms
step:278/1530 train_loss:4.0950 train_time:43842ms step_avg:163.59ms
step:279/1530 train_loss:4.0034 train_time:44008ms step_avg:163.60ms
step:280/1530 train_loss:4.1916 train_time:44176ms step_avg:163.61ms
step:281/1530 train_loss:3.9708 train_time:44344ms step_avg:163.63ms
step:282/1530 train_loss:3.9481 train_time:44511ms step_avg:163.65ms
step:283/1530 train_loss:3.9124 train_time:44678ms step_avg:163.65ms
step:284/1530 train_loss:4.0442 train_time:44844ms step_avg:163.66ms
step:285/1530 train_loss:4.0616 train_time:45009ms step_avg:163.67ms
step:286/1530 train_loss:4.0915 train_time:45173ms step_avg:163.67ms
step:287/1530 train_loss:3.9090 train_time:45339ms step_avg:163.68ms
step:288/1530 train_loss:4.0181 train_time:45504ms step_avg:163.69ms
step:289/1530 train_loss:3.8735 train_time:45670ms step_avg:163.69ms
step:290/1530 train_loss:3.8640 train_time:45835ms step_avg:163.70ms
step:291/1530 train_loss:3.9133 train_time:46001ms step_avg:163.70ms
step:292/1530 train_loss:3.8644 train_time:46168ms step_avg:163.72ms
step:293/1530 train_loss:3.9046 train_time:46333ms step_avg:163.72ms
step:294/1530 train_loss:3.9432 train_time:46498ms step_avg:163.72ms
step:295/1530 train_loss:3.8450 train_time:46662ms step_avg:163.73ms
step:296/1530 train_loss:3.8650 train_time:46828ms step_avg:163.73ms
step:297/1530 train_loss:3.8710 train_time:46993ms step_avg:163.74ms
step:298/1530 train_loss:3.9772 train_time:47158ms step_avg:163.74ms
step:299/1530 train_loss:3.8260 train_time:47323ms step_avg:163.75ms
step:300/1530 train_loss:3.9758 train_time:47488ms step_avg:163.75ms
step:301/1530 train_loss:3.9708 train_time:47653ms step_avg:163.76ms
step:302/1530 train_loss:3.9394 train_time:47818ms step_avg:163.76ms
step:303/1530 train_loss:3.9730 train_time:47983ms step_avg:163.77ms
step:304/1530 train_loss:3.9709 train_time:48149ms step_avg:163.77ms
step:305/1530 train_loss:4.4595 train_time:48313ms step_avg:163.77ms
step:306/1530 train_loss:3.9456 train_time:48477ms step_avg:163.77ms
step:307/1530 train_loss:3.8491 train_time:48644ms step_avg:163.79ms
step:308/1530 train_loss:3.9831 train_time:48809ms step_avg:163.79ms
step:309/1530 train_loss:3.8682 train_time:48974ms step_avg:163.79ms
step:310/1530 train_loss:4.0895 train_time:49139ms step_avg:163.80ms
step:311/1530 train_loss:3.9264 train_time:49305ms step_avg:163.80ms
step:312/1530 train_loss:3.8685 train_time:49470ms step_avg:163.81ms
step:313/1530 train_loss:3.9501 train_time:49636ms step_avg:163.82ms
step:314/1530 train_loss:4.0663 train_time:49802ms step_avg:163.82ms
step:315/1530 train_loss:3.9514 train_time:49967ms step_avg:163.83ms
step:316/1530 train_loss:3.7992 train_time:50132ms step_avg:163.83ms
step:317/1530 train_loss:3.8877 train_time:50296ms step_avg:163.83ms
step:318/1530 train_loss:3.9318 train_time:50461ms step_avg:163.83ms
step:319/1530 train_loss:3.8945 train_time:50627ms step_avg:163.84ms
step:320/1530 train_loss:4.0151 train_time:50792ms step_avg:163.84ms
step:321/1530 train_loss:3.9632 train_time:50956ms step_avg:163.85ms
step:322/1530 train_loss:3.9340 train_time:51122ms step_avg:163.85ms
step:323/1530 train_loss:4.0106 train_time:51287ms step_avg:163.86ms
step:324/1530 train_loss:3.9527 train_time:51453ms step_avg:163.86ms
step:325/1530 train_loss:4.0288 train_time:51619ms step_avg:163.87ms
step:326/1530 train_loss:3.8978 train_time:51785ms step_avg:163.88ms
step:327/1530 train_loss:4.4010 train_time:51950ms step_avg:163.88ms
step:328/1530 train_loss:4.0828 train_time:52115ms step_avg:163.88ms
step:329/1530 train_loss:3.8053 train_time:52280ms step_avg:163.89ms
step:330/1530 train_loss:3.7559 train_time:52446ms step_avg:163.89ms
step:331/1530 train_loss:3.9785 train_time:52611ms step_avg:163.90ms
step:332/1530 train_loss:3.9118 train_time:52776ms step_avg:163.90ms
step:333/1530 train_loss:3.8921 train_time:52941ms step_avg:163.90ms
step:334/1530 train_loss:3.8476 train_time:53106ms step_avg:163.91ms
step:335/1530 train_loss:4.0204 train_time:53271ms step_avg:163.91ms
step:336/1530 train_loss:3.9684 train_time:53437ms step_avg:163.92ms
step:337/1530 train_loss:4.4251 train_time:53602ms step_avg:163.92ms
step:338/1530 train_loss:3.9385 train_time:53767ms step_avg:163.92ms
step:339/1530 train_loss:3.8776 train_time:53933ms step_avg:163.93ms
step:340/1530 train_loss:3.9376 train_time:54098ms step_avg:163.93ms
step:341/1530 train_loss:3.8600 train_time:54265ms step_avg:163.94ms
step:342/1530 train_loss:3.8162 train_time:54432ms step_avg:163.95ms
step:343/1530 train_loss:3.8488 train_time:54601ms step_avg:163.97ms
step:344/1530 train_loss:3.9999 train_time:54769ms step_avg:163.98ms
step:345/1530 train_loss:3.8242 train_time:54939ms step_avg:164.00ms
step:346/1530 train_loss:3.7671 train_time:55108ms step_avg:164.01ms
step:347/1530 train_loss:3.8052 train_time:55274ms step_avg:164.02ms
step:348/1530 train_loss:3.8635 train_time:55443ms step_avg:164.03ms
step:349/1530 train_loss:3.8307 train_time:55611ms step_avg:164.04ms
step:350/1530 train_loss:3.5677 train_time:55780ms step_avg:164.06ms
step:351/1530 train_loss:3.8358 train_time:55949ms step_avg:164.07ms
step:352/1530 train_loss:4.1967 train_time:56117ms step_avg:164.08ms
step:353/1530 train_loss:3.6688 train_time:56285ms step_avg:164.10ms
step:354/1530 train_loss:3.9254 train_time:56452ms step_avg:164.10ms
step:355/1530 train_loss:3.7937 train_time:56622ms step_avg:164.12ms
step:356/1530 train_loss:3.8886 train_time:56790ms step_avg:164.13ms
step:357/1530 train_loss:3.7664 train_time:56958ms step_avg:164.14ms
step:358/1530 train_loss:3.8738 train_time:57127ms step_avg:164.16ms
step:359/1530 train_loss:3.7775 train_time:57295ms step_avg:164.17ms
step:360/1530 train_loss:3.4451 train_time:57465ms step_avg:164.19ms
step:361/1530 train_loss:4.0209 train_time:57633ms step_avg:164.20ms
step:362/1530 train_loss:3.9262 train_time:57801ms step_avg:164.21ms
step:363/1530 train_loss:3.8459 train_time:57968ms step_avg:164.22ms
step:364/1530 train_loss:3.7522 train_time:58136ms step_avg:164.23ms
step:365/1530 train_loss:3.9203 train_time:58305ms step_avg:164.24ms
step:366/1530 train_loss:3.8710 train_time:58473ms step_avg:164.25ms
step:367/1530 train_loss:3.8603 train_time:58641ms step_avg:164.26ms
step:368/1530 train_loss:3.8532 train_time:58808ms step_avg:164.27ms
step:369/1530 train_loss:3.7557 train_time:58976ms step_avg:164.28ms
step:370/1530 train_loss:3.8828 train_time:59145ms step_avg:164.29ms
step:371/1530 train_loss:3.7322 train_time:59312ms step_avg:164.30ms
step:372/1530 train_loss:3.6998 train_time:59481ms step_avg:164.31ms
step:373/1530 train_loss:3.9182 train_time:59650ms step_avg:164.32ms
step:374/1530 train_loss:3.8326 train_time:59816ms step_avg:164.33ms
step:375/1530 train_loss:3.8083 train_time:59985ms step_avg:164.34ms
step:375/1530 val_loss:3.8302 train_time:60034ms step_avg:164.48ms
step:376/1530 train_loss:3.8669 train_time:60155ms step_avg:164.36ms
step:377/1530 train_loss:3.7923 train_time:60456ms step_avg:164.73ms
step:378/1530 train_loss:3.8596 train_time:60634ms step_avg:164.77ms
step:379/1530 train_loss:3.8796 train_time:60957ms step_avg:165.20ms
step:380/1530 train_loss:3.9581 train_time:61124ms step_avg:165.20ms
step:381/1530 train_loss:3.8452 train_time:61293ms step_avg:165.21ms
step:382/1530 train_loss:3.8074 train_time:61462ms step_avg:165.22ms
step:383/1530 train_loss:3.8055 train_time:61630ms step_avg:165.23ms
step:384/1530 train_loss:3.8767 train_time:61798ms step_avg:165.23ms
step:385/1530 train_loss:3.8019 train_time:61965ms step_avg:165.24ms
step:386/1530 train_loss:3.9055 train_time:62132ms step_avg:165.25ms
step:387/1530 train_loss:4.0649 train_time:62300ms step_avg:165.25ms
step:388/1530 train_loss:3.7987 train_time:62467ms step_avg:165.26ms
step:389/1530 train_loss:3.8007 train_time:62635ms step_avg:165.26ms
step:390/1530 train_loss:3.9033 train_time:62803ms step_avg:165.27ms
step:391/1530 train_loss:3.8178 train_time:62971ms step_avg:165.28ms
step:392/1530 train_loss:3.9265 train_time:63138ms step_avg:165.28ms
step:393/1530 train_loss:3.7675 train_time:63306ms step_avg:165.29ms
step:394/1530 train_loss:3.8863 train_time:63474ms step_avg:165.30ms
step:395/1530 train_loss:3.6323 train_time:63641ms step_avg:165.30ms
step:396/1530 train_loss:3.8421 train_time:63808ms step_avg:165.31ms
step:397/1530 train_loss:3.8650 train_time:63978ms step_avg:165.32ms
step:398/1530 train_loss:3.8905 train_time:64144ms step_avg:165.32ms
step:399/1530 train_loss:3.7777 train_time:64309ms step_avg:165.32ms
step:400/1530 train_loss:3.8363 train_time:64479ms step_avg:165.33ms
step:401/1530 train_loss:3.9216 train_time:64647ms step_avg:165.34ms
step:402/1530 train_loss:3.8495 train_time:64814ms step_avg:165.34ms
step:403/1530 train_loss:3.9717 train_time:64981ms step_avg:165.35ms
step:404/1530 train_loss:3.6832 train_time:65147ms step_avg:165.35ms
step:405/1530 train_loss:3.7934 train_time:65316ms step_avg:165.36ms
step:406/1530 train_loss:4.1009 train_time:65482ms step_avg:165.36ms
step:407/1530 train_loss:3.7798 train_time:65649ms step_avg:165.36ms
step:408/1530 train_loss:3.8292 train_time:65816ms step_avg:165.37ms
step:409/1530 train_loss:3.8596 train_time:65983ms step_avg:165.37ms
step:410/1530 train_loss:3.7620 train_time:66150ms step_avg:165.37ms
step:411/1530 train_loss:3.7667 train_time:66318ms step_avg:165.38ms
step:412/1530 train_loss:4.1886 train_time:66484ms step_avg:165.38ms
step:413/1530 train_loss:4.0281 train_time:66651ms step_avg:165.39ms
step:414/1530 train_loss:4.0191 train_time:66818ms step_avg:165.39ms
step:415/1530 train_loss:3.7571 train_time:66985ms step_avg:165.39ms
step:416/1530 train_loss:3.7722 train_time:67151ms step_avg:165.40ms
step:417/1530 train_loss:3.9586 train_time:67319ms step_avg:165.40ms
step:418/1530 train_loss:3.7024 train_time:67486ms step_avg:165.41ms
step:419/1530 train_loss:3.8147 train_time:67652ms step_avg:165.41ms
step:420/1530 train_loss:3.7087 train_time:67819ms step_avg:165.41ms
step:421/1530 train_loss:3.6554 train_time:67986ms step_avg:165.42ms
step:422/1530 train_loss:3.7939 train_time:68153ms step_avg:165.42ms
step:423/1530 train_loss:3.8791 train_time:68320ms step_avg:165.42ms
step:424/1530 train_loss:3.6190 train_time:68487ms step_avg:165.43ms
step:425/1530 train_loss:3.8049 train_time:68655ms step_avg:165.43ms
step:426/1530 train_loss:3.6776 train_time:68822ms step_avg:165.44ms
step:427/1530 train_loss:3.8970 train_time:68990ms step_avg:165.44ms
step:428/1530 train_loss:3.8200 train_time:69157ms step_avg:165.45ms
step:429/1530 train_loss:3.7707 train_time:69324ms step_avg:165.45ms
step:430/1530 train_loss:3.7128 train_time:69492ms step_avg:165.46ms
step:431/1530 train_loss:3.6335 train_time:69659ms step_avg:165.46ms
step:432/1530 train_loss:3.7775 train_time:69826ms step_avg:165.46ms
step:433/1530 train_loss:3.8248 train_time:69994ms step_avg:165.47ms
step:434/1530 train_loss:3.7771 train_time:70161ms step_avg:165.47ms
step:435/1530 train_loss:3.8133 train_time:70328ms step_avg:165.48ms
step:436/1530 train_loss:3.8350 train_time:70497ms step_avg:165.49ms
step:437/1530 train_loss:3.7283 train_time:70664ms step_avg:165.49ms
step:438/1530 train_loss:3.7037 train_time:70830ms step_avg:165.49ms
step:439/1530 train_loss:3.7207 train_time:70999ms step_avg:165.50ms
step:440/1530 train_loss:3.8897 train_time:71165ms step_avg:165.50ms
step:441/1530 train_loss:3.7669 train_time:71332ms step_avg:165.50ms
step:442/1530 train_loss:3.7484 train_time:71500ms step_avg:165.51ms
step:443/1530 train_loss:3.6293 train_time:71667ms step_avg:165.51ms
step:444/1530 train_loss:3.9305 train_time:71834ms step_avg:165.52ms
step:445/1530 train_loss:3.8472 train_time:72001ms step_avg:165.52ms
step:446/1530 train_loss:3.8407 train_time:72169ms step_avg:165.53ms
step:447/1530 train_loss:3.7556 train_time:72337ms step_avg:165.53ms
step:448/1530 train_loss:3.8514 train_time:72503ms step_avg:165.53ms
step:449/1530 train_loss:3.6922 train_time:72671ms step_avg:165.54ms
step:450/1530 train_loss:3.7246 train_time:72839ms step_avg:165.54ms
step:451/1530 train_loss:3.5865 train_time:73005ms step_avg:165.55ms
step:452/1530 train_loss:3.7189 train_time:73173ms step_avg:165.55ms
step:453/1530 train_loss:3.6766 train_time:73340ms step_avg:165.55ms
step:454/1530 train_loss:3.6474 train_time:73506ms step_avg:165.56ms
step:455/1530 train_loss:3.8444 train_time:73675ms step_avg:165.56ms
step:456/1530 train_loss:3.7278 train_time:73843ms step_avg:165.57ms
step:457/1530 train_loss:3.7819 train_time:74013ms step_avg:165.58ms
step:458/1530 train_loss:3.8356 train_time:74182ms step_avg:165.59ms
step:459/1530 train_loss:3.6377 train_time:74353ms step_avg:165.60ms
step:460/1530 train_loss:3.7916 train_time:74523ms step_avg:165.61ms
step:461/1530 train_loss:3.6992 train_time:74692ms step_avg:165.61ms
step:462/1530 train_loss:3.7392 train_time:74862ms step_avg:165.62ms
step:463/1530 train_loss:3.7837 train_time:75034ms step_avg:165.64ms
step:464/1530 train_loss:3.7224 train_time:75204ms step_avg:165.65ms
step:465/1530 train_loss:3.7269 train_time:75373ms step_avg:165.65ms
step:466/1530 train_loss:3.8093 train_time:75542ms step_avg:165.66ms
step:467/1530 train_loss:3.8352 train_time:75713ms step_avg:165.67ms
step:468/1530 train_loss:3.7949 train_time:75882ms step_avg:165.68ms
step:469/1530 train_loss:3.6906 train_time:76053ms step_avg:165.69ms
step:470/1530 train_loss:3.7740 train_time:76222ms step_avg:165.70ms
step:471/1530 train_loss:3.8188 train_time:76393ms step_avg:165.71ms
step:472/1530 train_loss:3.7871 train_time:76564ms step_avg:165.72ms
step:473/1530 train_loss:3.7172 train_time:76734ms step_avg:165.73ms
step:474/1530 train_loss:3.5930 train_time:76904ms step_avg:165.74ms
step:475/1530 train_loss:4.0361 train_time:77074ms step_avg:165.75ms
step:476/1530 train_loss:3.7586 train_time:77243ms step_avg:165.76ms
step:477/1530 train_loss:3.5990 train_time:77413ms step_avg:165.77ms
step:478/1530 train_loss:3.8291 train_time:77583ms step_avg:165.77ms
step:479/1530 train_loss:3.7782 train_time:77751ms step_avg:165.78ms
step:480/1530 train_loss:3.9267 train_time:77922ms step_avg:165.79ms
step:481/1530 train_loss:3.7340 train_time:78093ms step_avg:165.80ms
step:482/1530 train_loss:3.5320 train_time:78263ms step_avg:165.81ms
step:483/1530 train_loss:3.8080 train_time:78432ms step_avg:165.82ms
step:484/1530 train_loss:3.6654 train_time:78604ms step_avg:165.83ms
step:485/1530 train_loss:3.6622 train_time:78774ms step_avg:165.84ms
step:486/1530 train_loss:3.5806 train_time:78944ms step_avg:165.85ms
step:487/1530 train_loss:3.6953 train_time:79113ms step_avg:165.85ms
step:488/1530 train_loss:3.8787 train_time:79283ms step_avg:165.86ms
step:489/1530 train_loss:3.7155 train_time:79454ms step_avg:165.87ms
step:490/1530 train_loss:3.5944 train_time:79623ms step_avg:165.88ms
step:491/1530 train_loss:3.6254 train_time:79793ms step_avg:165.89ms
step:492/1530 train_loss:3.7362 train_time:79963ms step_avg:165.90ms
step:493/1530 train_loss:3.5834 train_time:80134ms step_avg:165.91ms
step:494/1530 train_loss:3.7070 train_time:80305ms step_avg:165.92ms
step:495/1530 train_loss:3.6641 train_time:80476ms step_avg:165.93ms
step:496/1530 train_loss:3.5111 train_time:80645ms step_avg:165.94ms
step:497/1530 train_loss:3.7333 train_time:80813ms step_avg:165.94ms
step:498/1530 train_loss:3.7881 train_time:80983ms step_avg:165.95ms
step:499/1530 train_loss:3.8276 train_time:81153ms step_avg:165.96ms
step:500/1530 train_loss:3.7368 train_time:81323ms step_avg:165.96ms
step:500/1530 val_loss:3.7113 train_time:81372ms step_avg:166.06ms
step:501/1530 train_loss:3.8114 train_time:81495ms step_avg:165.98ms
step:502/1530 train_loss:3.7552 train_time:81665ms step_avg:165.99ms
step:503/1530 train_loss:3.7787 train_time:81836ms step_avg:166.00ms
step:504/1530 train_loss:3.7295 train_time:82004ms step_avg:166.00ms
step:505/1530 train_loss:3.8134 train_time:82174ms step_avg:166.01ms
step:506/1530 train_loss:3.6688 train_time:82343ms step_avg:166.01ms
step:507/1530 train_loss:3.7701 train_time:82511ms step_avg:166.02ms
step:508/1530 train_loss:3.8280 train_time:82681ms step_avg:166.03ms
step:509/1530 train_loss:3.7824 train_time:82849ms step_avg:166.03ms
step:510/1530 train_loss:3.5865 train_time:83020ms step_avg:166.04ms
step:511/1530 train_loss:3.7778 train_time:83189ms step_avg:166.05ms
step:512/1530 train_loss:3.7262 train_time:83360ms step_avg:166.06ms
step:513/1530 train_loss:3.6699 train_time:83529ms step_avg:166.06ms
step:514/1530 train_loss:3.8134 train_time:83700ms step_avg:166.07ms
step:515/1530 train_loss:3.7418 train_time:83869ms step_avg:166.08ms
step:516/1530 train_loss:4.0897 train_time:84040ms step_avg:166.09ms
step:517/1530 train_loss:3.6931 train_time:84208ms step_avg:166.09ms
step:518/1530 train_loss:3.7812 train_time:84377ms step_avg:166.10ms
step:519/1530 train_loss:3.6639 train_time:84546ms step_avg:166.10ms
step:520/1530 train_loss:3.6855 train_time:84717ms step_avg:166.11ms
step:521/1530 train_loss:3.6629 train_time:84885ms step_avg:166.12ms
step:522/1530 train_loss:3.6616 train_time:85056ms step_avg:166.13ms
step:523/1530 train_loss:4.3038 train_time:85225ms step_avg:166.13ms
step:524/1530 train_loss:3.7428 train_time:85394ms step_avg:166.14ms
step:525/1530 train_loss:3.6872 train_time:85563ms step_avg:166.14ms
step:526/1530 train_loss:3.7101 train_time:85731ms step_avg:166.14ms
step:527/1530 train_loss:3.6681 train_time:85901ms step_avg:166.15ms
step:528/1530 train_loss:3.6357 train_time:86071ms step_avg:166.16ms
step:529/1530 train_loss:3.8518 train_time:86240ms step_avg:166.17ms
step:530/1530 train_loss:3.6468 train_time:86409ms step_avg:166.17ms
step:531/1530 train_loss:3.9189 train_time:86578ms step_avg:166.18ms
step:532/1530 train_loss:3.7391 train_time:86747ms step_avg:166.18ms
step:533/1530 train_loss:3.6560 train_time:86916ms step_avg:166.19ms
step:534/1530 train_loss:3.6728 train_time:87084ms step_avg:166.19ms
step:535/1530 train_loss:3.6124 train_time:87253ms step_avg:166.20ms
step:536/1530 train_loss:3.7559 train_time:87424ms step_avg:166.20ms
step:537/1530 train_loss:3.7300 train_time:87593ms step_avg:166.21ms
step:538/1530 train_loss:3.6298 train_time:87763ms step_avg:166.22ms
step:539/1530 train_loss:4.1214 train_time:87936ms step_avg:166.23ms
step:540/1530 train_loss:3.6798 train_time:88104ms step_avg:166.23ms
step:541/1530 train_loss:3.7872 train_time:88274ms step_avg:166.24ms
step:542/1530 train_loss:3.5902 train_time:88442ms step_avg:166.24ms
step:543/1530 train_loss:3.5890 train_time:88612ms step_avg:166.25ms
step:544/1530 train_loss:3.6437 train_time:88780ms step_avg:166.25ms
step:545/1530 train_loss:3.5953 train_time:88947ms step_avg:166.26ms
step:546/1530 train_loss:3.6313 train_time:89119ms step_avg:166.27ms
step:547/1530 train_loss:3.6462 train_time:89288ms step_avg:166.27ms
step:548/1530 train_loss:3.6189 train_time:89458ms step_avg:166.28ms
step:549/1530 train_loss:3.7266 train_time:89626ms step_avg:166.28ms
step:550/1530 train_loss:3.6286 train_time:89796ms step_avg:166.29ms
step:551/1530 train_loss:3.6308 train_time:89965ms step_avg:166.29ms
step:552/1530 train_loss:3.9400 train_time:90133ms step_avg:166.30ms
step:553/1530 train_loss:3.7597 train_time:90304ms step_avg:166.31ms
step:554/1530 train_loss:3.7164 train_time:90472ms step_avg:166.31ms
step:555/1530 train_loss:3.6277 train_time:90642ms step_avg:166.31ms
step:556/1530 train_loss:3.7020 train_time:90810ms step_avg:166.32ms
step:557/1530 train_loss:3.3176 train_time:90979ms step_avg:166.32ms
step:558/1530 train_loss:3.6146 train_time:91147ms step_avg:166.33ms
step:559/1530 train_loss:3.6566 train_time:91316ms step_avg:166.33ms
step:560/1530 train_loss:3.6968 train_time:91485ms step_avg:166.34ms
step:561/1530 train_loss:3.6209 train_time:91653ms step_avg:166.34ms
step:562/1530 train_loss:3.5543 train_time:91822ms step_avg:166.34ms
step:563/1530 train_loss:3.7543 train_time:91991ms step_avg:166.35ms
step:564/1530 train_loss:3.5756 train_time:92161ms step_avg:166.35ms
step:565/1530 train_loss:3.6807 train_time:92330ms step_avg:166.36ms
step:566/1530 train_loss:3.6270 train_time:92631ms step_avg:166.60ms
step:567/1530 train_loss:3.6037 train_time:92809ms step_avg:166.62ms
step:568/1530 train_loss:3.6903 train_time:92979ms step_avg:166.63ms
step:569/1530 train_loss:3.6522 train_time:93302ms step_avg:166.91ms
step:570/1530 train_loss:3.6974 train_time:93475ms step_avg:166.92ms
step:571/1530 train_loss:3.7597 train_time:93645ms step_avg:166.93ms
step:572/1530 train_loss:3.7290 train_time:93818ms step_avg:166.94ms
step:573/1530 train_loss:3.7432 train_time:93988ms step_avg:166.94ms
step:574/1530 train_loss:3.7803 train_time:94163ms step_avg:166.96ms
step:575/1530 train_loss:3.7347 train_time:94335ms step_avg:166.97ms
step:576/1530 train_loss:3.7615 train_time:94506ms step_avg:166.97ms
step:577/1530 train_loss:3.6767 train_time:94677ms step_avg:166.98ms
step:578/1530 train_loss:3.6712 train_time:94847ms step_avg:166.98ms
step:579/1530 train_loss:3.6720 train_time:95019ms step_avg:166.99ms
step:580/1530 train_loss:3.5931 train_time:95189ms step_avg:167.00ms
step:581/1530 train_loss:3.6418 train_time:95360ms step_avg:167.01ms
step:582/1530 train_loss:3.8473 train_time:95529ms step_avg:167.01ms
step:583/1530 train_loss:3.6246 train_time:95702ms step_avg:167.02ms
step:584/1530 train_loss:3.6046 train_time:95875ms step_avg:167.03ms
step:585/1530 train_loss:3.7955 train_time:96044ms step_avg:167.03ms
step:586/1530 train_loss:3.5196 train_time:96217ms step_avg:167.04ms
step:587/1530 train_loss:3.6705 train_time:96387ms step_avg:167.05ms
step:588/1530 train_loss:3.6443 train_time:96557ms step_avg:167.05ms
step:589/1530 train_loss:4.0023 train_time:96729ms step_avg:167.06ms
step:590/1530 train_loss:3.7836 train_time:96901ms step_avg:167.07ms
step:591/1530 train_loss:3.5162 train_time:97072ms step_avg:167.08ms
step:592/1530 train_loss:3.5336 train_time:97245ms step_avg:167.09ms
step:593/1530 train_loss:3.5085 train_time:97420ms step_avg:167.10ms
step:594/1530 train_loss:3.5630 train_time:97591ms step_avg:167.11ms
step:595/1530 train_loss:3.9158 train_time:97764ms step_avg:167.12ms
step:596/1530 train_loss:3.6518 train_time:97937ms step_avg:167.13ms
step:597/1530 train_loss:3.5887 train_time:98107ms step_avg:167.13ms
step:598/1530 train_loss:3.6622 train_time:98278ms step_avg:167.14ms
step:599/1530 train_loss:3.4790 train_time:98447ms step_avg:167.14ms
step:600/1530 train_loss:3.5973 train_time:98619ms step_avg:167.15ms
step:601/1530 train_loss:3.6520 train_time:98790ms step_avg:167.16ms
step:602/1530 train_loss:3.6768 train_time:98963ms step_avg:167.17ms
step:603/1530 train_loss:3.7877 train_time:99135ms step_avg:167.18ms
step:604/1530 train_loss:3.6146 train_time:99306ms step_avg:167.18ms
step:605/1530 train_loss:3.6161 train_time:99478ms step_avg:167.19ms
step:606/1530 train_loss:3.5739 train_time:99651ms step_avg:167.20ms
step:607/1530 train_loss:3.8436 train_time:99824ms step_avg:167.21ms
step:608/1530 train_loss:3.6318 train_time:99996ms step_avg:167.22ms
step:609/1530 train_loss:3.6240 train_time:100167ms step_avg:167.22ms
step:610/1530 train_loss:3.7074 train_time:100337ms step_avg:167.23ms
step:611/1530 train_loss:3.6023 train_time:100508ms step_avg:167.24ms
step:612/1530 train_loss:3.5690 train_time:100680ms step_avg:167.24ms
step:613/1530 train_loss:3.7685 train_time:100849ms step_avg:167.25ms
step:614/1530 train_loss:3.6949 train_time:101023ms step_avg:167.26ms
step:615/1530 train_loss:3.6917 train_time:101195ms step_avg:167.26ms
step:616/1530 train_loss:3.6369 train_time:101364ms step_avg:167.27ms
step:617/1530 train_loss:3.5606 train_time:101538ms step_avg:167.28ms
step:618/1530 train_loss:3.6900 train_time:101707ms step_avg:167.28ms
step:619/1530 train_loss:3.5502 train_time:101879ms step_avg:167.29ms
step:620/1530 train_loss:3.5903 train_time:102049ms step_avg:167.29ms
step:621/1530 train_loss:3.9257 train_time:102222ms step_avg:167.30ms
step:622/1530 train_loss:3.5748 train_time:102393ms step_avg:167.31ms
step:623/1530 train_loss:3.6087 train_time:102566ms step_avg:167.32ms
step:624/1530 train_loss:3.6945 train_time:102737ms step_avg:167.32ms
step:625/1530 train_loss:3.7007 train_time:102908ms step_avg:167.33ms
step:625/1530 val_loss:3.6259 train_time:102957ms step_avg:167.41ms
step:626/1530 train_loss:3.7417 train_time:103081ms step_avg:167.34ms
step:627/1530 train_loss:3.7121 train_time:103255ms step_avg:167.35ms
step:628/1530 train_loss:3.7641 train_time:103424ms step_avg:167.35ms
step:629/1530 train_loss:3.5921 train_time:103596ms step_avg:167.36ms
step:630/1530 train_loss:3.7323 train_time:103767ms step_avg:167.37ms
step:631/1530 train_loss:3.7381 train_time:103937ms step_avg:167.37ms
step:632/1530 train_loss:3.6514 train_time:104108ms step_avg:167.38ms
step:633/1530 train_loss:3.6085 train_time:104278ms step_avg:167.38ms
step:634/1530 train_loss:3.7032 train_time:104450ms step_avg:167.39ms
step:635/1530 train_loss:3.9590 train_time:104621ms step_avg:167.39ms
step:636/1530 train_loss:3.5547 train_time:104792ms step_avg:167.40ms
step:637/1530 train_loss:3.3628 train_time:104964ms step_avg:167.41ms
step:638/1530 train_loss:3.5919 train_time:105133ms step_avg:167.41ms
step:639/1530 train_loss:3.6324 train_time:105303ms step_avg:167.41ms
step:640/1530 train_loss:3.5725 train_time:105474ms step_avg:167.42ms
step:641/1530 train_loss:3.5894 train_time:105644ms step_avg:167.42ms
step:642/1530 train_loss:3.6362 train_time:105814ms step_avg:167.43ms
step:643/1530 train_loss:3.5942 train_time:105985ms step_avg:167.43ms
step:644/1530 train_loss:3.5694 train_time:106155ms step_avg:167.44ms
step:645/1530 train_loss:3.7784 train_time:106327ms step_avg:167.44ms
step:646/1530 train_loss:3.6755 train_time:106497ms step_avg:167.45ms
step:647/1530 train_loss:3.6717 train_time:106669ms step_avg:167.45ms
step:648/1530 train_loss:3.7145 train_time:106840ms step_avg:167.46ms
step:649/1530 train_loss:3.7681 train_time:107011ms step_avg:167.47ms
step:650/1530 train_loss:3.6225 train_time:107181ms step_avg:167.47ms
step:651/1530 train_loss:3.7648 train_time:107354ms step_avg:167.48ms
step:652/1530 train_loss:3.5837 train_time:107524ms step_avg:167.48ms
step:653/1530 train_loss:3.6607 train_time:107694ms step_avg:167.49ms
step:654/1530 train_loss:3.4321 train_time:107866ms step_avg:167.49ms
step:655/1530 train_loss:3.5863 train_time:108036ms step_avg:167.50ms
step:656/1530 train_loss:3.5759 train_time:108205ms step_avg:167.50ms
step:657/1530 train_loss:3.5015 train_time:108376ms step_avg:167.50ms
step:658/1530 train_loss:3.6897 train_time:108544ms step_avg:167.51ms
step:659/1530 train_loss:3.5854 train_time:108715ms step_avg:167.51ms
step:660/1530 train_loss:3.6811 train_time:108885ms step_avg:167.52ms
step:661/1530 train_loss:3.7600 train_time:109056ms step_avg:167.52ms
step:662/1530 train_loss:3.6728 train_time:109226ms step_avg:167.52ms
step:663/1530 train_loss:3.5552 train_time:109396ms step_avg:167.53ms
step:664/1530 train_loss:3.6156 train_time:109567ms step_avg:167.53ms
step:665/1530 train_loss:3.5031 train_time:109739ms step_avg:167.54ms
step:666/1530 train_loss:3.7867 train_time:109908ms step_avg:167.54ms
step:667/1530 train_loss:3.6069 train_time:110079ms step_avg:167.55ms
step:668/1530 train_loss:3.6505 train_time:110250ms step_avg:167.55ms
step:669/1530 train_loss:3.4928 train_time:110420ms step_avg:167.56ms
step:670/1530 train_loss:3.5989 train_time:110590ms step_avg:167.56ms
step:671/1530 train_loss:3.5643 train_time:110760ms step_avg:167.56ms
step:672/1530 train_loss:3.5698 train_time:110933ms step_avg:167.57ms
step:673/1530 train_loss:3.8526 train_time:111103ms step_avg:167.58ms
step:674/1530 train_loss:3.6237 train_time:111274ms step_avg:167.58ms
step:675/1530 train_loss:3.7153 train_time:111446ms step_avg:167.59ms
step:676/1530 train_loss:3.4911 train_time:111616ms step_avg:167.59ms
step:677/1530 train_loss:3.6036 train_time:111788ms step_avg:167.60ms
step:678/1530 train_loss:3.5600 train_time:111959ms step_avg:167.60ms
step:679/1530 train_loss:3.6851 train_time:112130ms step_avg:167.61ms
step:680/1530 train_loss:3.5844 train_time:112300ms step_avg:167.61ms
step:681/1530 train_loss:3.6139 train_time:112473ms step_avg:167.62ms
step:682/1530 train_loss:3.6739 train_time:112648ms step_avg:167.63ms
step:683/1530 train_loss:3.7426 train_time:112820ms step_avg:167.64ms
step:684/1530 train_loss:3.6520 train_time:112992ms step_avg:167.64ms
step:685/1530 train_loss:3.6887 train_time:113165ms step_avg:167.65ms
step:686/1530 train_loss:3.6385 train_time:113338ms step_avg:167.66ms
step:687/1530 train_loss:3.6707 train_time:113511ms step_avg:167.67ms
step:688/1530 train_loss:3.2162 train_time:113685ms step_avg:167.68ms
step:689/1530 train_loss:3.4125 train_time:113859ms step_avg:167.69ms
step:690/1530 train_loss:3.5418 train_time:114034ms step_avg:167.70ms
step:691/1530 train_loss:3.4111 train_time:114207ms step_avg:167.70ms
step:692/1530 train_loss:3.6322 train_time:114379ms step_avg:167.71ms
step:693/1530 train_loss:3.6491 train_time:114553ms step_avg:167.72ms
step:694/1530 train_loss:3.5574 train_time:114726ms step_avg:167.73ms
step:695/1530 train_loss:3.5356 train_time:114896ms step_avg:167.73ms
step:696/1530 train_loss:3.8610 train_time:115071ms step_avg:167.74ms
step:697/1530 train_loss:3.5858 train_time:115243ms step_avg:167.75ms
step:698/1530 train_loss:3.6487 train_time:115415ms step_avg:167.75ms
step:699/1530 train_loss:3.7675 train_time:115589ms step_avg:167.76ms
step:700/1530 train_loss:3.5725 train_time:115762ms step_avg:167.77ms
step:701/1530 train_loss:3.5435 train_time:115934ms step_avg:167.78ms
step:702/1530 train_loss:3.5114 train_time:116108ms step_avg:167.79ms
step:703/1530 train_loss:3.5035 train_time:116279ms step_avg:167.79ms
step:704/1530 train_loss:3.5765 train_time:116453ms step_avg:167.80ms
step:705/1530 train_loss:3.5664 train_time:116630ms step_avg:167.81ms
step:706/1530 train_loss:3.5765 train_time:116804ms step_avg:167.82ms
step:707/1530 train_loss:3.6488 train_time:116979ms step_avg:167.83ms
step:708/1530 train_loss:3.6086 train_time:117152ms step_avg:167.84ms
step:709/1530 train_loss:3.5827 train_time:117325ms step_avg:167.85ms
step:710/1530 train_loss:3.5403 train_time:117497ms step_avg:167.85ms
step:711/1530 train_loss:3.5981 train_time:117670ms step_avg:167.86ms
step:712/1530 train_loss:3.6492 train_time:117845ms step_avg:167.87ms
step:713/1530 train_loss:3.6498 train_time:118019ms step_avg:167.88ms
step:714/1530 train_loss:3.5645 train_time:118191ms step_avg:167.89ms
step:715/1530 train_loss:3.5727 train_time:118363ms step_avg:167.89ms
step:716/1530 train_loss:3.5973 train_time:118535ms step_avg:167.90ms
step:717/1530 train_loss:3.7078 train_time:118710ms step_avg:167.91ms
step:718/1530 train_loss:3.6022 train_time:118882ms step_avg:167.91ms
step:719/1530 train_loss:3.6769 train_time:119055ms step_avg:167.92ms
step:720/1530 train_loss:3.8413 train_time:119230ms step_avg:167.93ms
step:721/1530 train_loss:3.4623 train_time:119403ms step_avg:167.94ms
step:722/1530 train_loss:3.7392 train_time:119576ms step_avg:167.94ms
step:723/1530 train_loss:3.7720 train_time:119747ms step_avg:167.95ms
step:724/1530 train_loss:3.5697 train_time:119921ms step_avg:167.96ms
step:725/1530 train_loss:3.6560 train_time:120094ms step_avg:167.96ms
step:726/1530 train_loss:3.5320 train_time:120267ms step_avg:167.97ms
step:727/1530 train_loss:3.5838 train_time:120442ms step_avg:167.98ms
step:728/1530 train_loss:3.7332 train_time:120614ms step_avg:167.99ms
step:729/1530 train_loss:3.6744 train_time:120786ms step_avg:167.99ms
step:730/1530 train_loss:3.6683 train_time:120960ms step_avg:168.00ms
step:731/1530 train_loss:3.5540 train_time:121134ms step_avg:168.01ms
step:732/1530 train_loss:3.6015 train_time:121305ms step_avg:168.01ms
step:733/1530 train_loss:3.8316 train_time:121479ms step_avg:168.02ms
step:734/1530 train_loss:3.5653 train_time:121655ms step_avg:168.03ms
step:735/1530 train_loss:3.6193 train_time:121827ms step_avg:168.04ms
step:736/1530 train_loss:3.7417 train_time:121999ms step_avg:168.04ms
step:737/1530 train_loss:3.6791 train_time:122173ms step_avg:168.05ms
step:738/1530 train_loss:3.6047 train_time:122344ms step_avg:168.06ms
step:739/1530 train_loss:3.5008 train_time:122515ms step_avg:168.06ms
step:740/1530 train_loss:4.1131 train_time:122691ms step_avg:168.07ms
step:741/1530 train_loss:3.4943 train_time:122862ms step_avg:168.07ms
step:742/1530 train_loss:3.5621 train_time:123035ms step_avg:168.08ms
step:743/1530 train_loss:3.5806 train_time:123210ms step_avg:168.09ms
step:744/1530 train_loss:3.6450 train_time:123383ms step_avg:168.10ms
step:745/1530 train_loss:3.5902 train_time:123557ms step_avg:168.10ms
step:746/1530 train_loss:3.5938 train_time:123729ms step_avg:168.11ms
step:747/1530 train_loss:3.6553 train_time:123902ms step_avg:168.12ms
step:748/1530 train_loss:3.5676 train_time:124079ms step_avg:168.13ms
step:749/1530 train_loss:3.5680 train_time:124251ms step_avg:168.13ms
step:750/1530 train_loss:3.6037 train_time:124422ms step_avg:168.14ms
step:750/1530 val_loss:3.5679 train_time:124470ms step_avg:168.20ms
step:751/1530 train_loss:3.5703 train_time:124594ms step_avg:168.14ms
step:752/1530 train_loss:3.6153 train_time:124766ms step_avg:168.15ms
step:753/1530 train_loss:3.6219 train_time:124940ms step_avg:168.16ms
step:754/1530 train_loss:3.5970 train_time:125111ms step_avg:168.16ms
step:755/1530 train_loss:3.6814 train_time:125417ms step_avg:168.34ms
step:756/1530 train_loss:3.4635 train_time:125601ms step_avg:168.37ms
step:757/1530 train_loss:3.7273 train_time:125774ms step_avg:168.37ms
step:758/1530 train_loss:3.6550 train_time:125944ms step_avg:168.37ms
step:759/1530 train_loss:3.5961 train_time:126269ms step_avg:168.58ms
step:760/1530 train_loss:3.7050 train_time:126441ms step_avg:168.59ms
step:761/1530 train_loss:3.4059 train_time:126612ms step_avg:168.59ms
step:762/1530 train_loss:3.5542 train_time:126785ms step_avg:168.60ms
step:763/1530 train_loss:3.6632 train_time:126958ms step_avg:168.60ms
step:764/1530 train_loss:3.3235 train_time:127131ms step_avg:168.61ms
step:765/1530 train_loss:3.7366 train_time:127304ms step_avg:168.61ms
step:766/1530 train_loss:3.5780 train_time:127478ms step_avg:168.62ms
step:767/1530 train_loss:3.5697 train_time:127649ms step_avg:168.62ms
step:768/1530 train_loss:3.5743 train_time:127824ms step_avg:168.63ms
step:769/1530 train_loss:3.5865 train_time:127997ms step_avg:168.64ms
step:770/1530 train_loss:3.6437 train_time:128169ms step_avg:168.64ms
step:771/1530 train_loss:3.8843 train_time:128342ms step_avg:168.65ms
step:772/1530 train_loss:3.4559 train_time:128513ms step_avg:168.65ms
step:773/1530 train_loss:3.6305 train_time:128685ms step_avg:168.66ms
step:774/1530 train_loss:3.6404 train_time:128858ms step_avg:168.66ms
step:775/1530 train_loss:3.6080 train_time:129029ms step_avg:168.67ms
step:776/1530 train_loss:3.3973 train_time:129204ms step_avg:168.67ms
step:777/1530 train_loss:3.3966 train_time:129379ms step_avg:168.68ms
step:778/1530 train_loss:3.4877 train_time:129549ms step_avg:168.68ms
step:779/1530 train_loss:3.5845 train_time:129721ms step_avg:168.69ms
step:780/1530 train_loss:3.5872 train_time:129893ms step_avg:168.69ms
step:781/1530 train_loss:3.6746 train_time:130065ms step_avg:168.70ms
step:782/1530 train_loss:3.5881 train_time:130236ms step_avg:168.70ms
step:783/1530 train_loss:3.5646 train_time:130407ms step_avg:168.70ms
step:784/1530 train_loss:3.6024 train_time:130581ms step_avg:168.71ms
step:785/1530 train_loss:3.5645 train_time:130753ms step_avg:168.71ms
step:786/1530 train_loss:3.4439 train_time:130926ms step_avg:168.72ms
step:787/1530 train_loss:3.7315 train_time:131099ms step_avg:168.73ms
step:788/1530 train_loss:3.5044 train_time:131271ms step_avg:168.73ms
step:789/1530 train_loss:3.5472 train_time:131442ms step_avg:168.73ms
step:790/1530 train_loss:3.6311 train_time:131616ms step_avg:168.74ms
step:791/1530 train_loss:3.7717 train_time:131791ms step_avg:168.75ms
step:792/1530 train_loss:3.7576 train_time:131963ms step_avg:168.75ms
step:793/1530 train_loss:3.4625 train_time:132134ms step_avg:168.75ms
step:794/1530 train_loss:3.5936 train_time:132308ms step_avg:168.76ms
step:795/1530 train_loss:3.6787 train_time:132481ms step_avg:168.77ms
step:796/1530 train_loss:3.7499 train_time:132657ms step_avg:168.78ms
step:797/1530 train_loss:3.5272 train_time:132830ms step_avg:168.78ms
step:798/1530 train_loss:3.6485 train_time:133006ms step_avg:168.79ms
step:799/1530 train_loss:3.5344 train_time:133184ms step_avg:168.80ms
step:800/1530 train_loss:3.5317 train_time:133357ms step_avg:168.81ms
step:801/1530 train_loss:3.6351 train_time:133530ms step_avg:168.81ms
step:802/1530 train_loss:3.5004 train_time:133708ms step_avg:168.82ms
step:803/1530 train_loss:3.4908 train_time:133880ms step_avg:168.83ms
step:804/1530 train_loss:3.6286 train_time:134055ms step_avg:168.83ms
step:805/1530 train_loss:3.5202 train_time:134230ms step_avg:168.84ms
step:806/1530 train_loss:3.5630 train_time:134405ms step_avg:168.85ms
step:807/1530 train_loss:3.6427 train_time:134578ms step_avg:168.86ms
step:808/1530 train_loss:3.5424 train_time:134754ms step_avg:168.86ms
step:809/1530 train_loss:3.4921 train_time:134927ms step_avg:168.87ms
step:810/1530 train_loss:3.5661 train_time:135101ms step_avg:168.88ms
step:811/1530 train_loss:3.5793 train_time:135273ms step_avg:168.88ms
step:812/1530 train_loss:3.6040 train_time:135446ms step_avg:168.89ms
step:813/1530 train_loss:3.6270 train_time:135619ms step_avg:168.89ms
step:814/1530 train_loss:3.5628 train_time:135792ms step_avg:168.90ms
step:815/1530 train_loss:3.5705 train_time:135966ms step_avg:168.90ms
step:816/1530 train_loss:3.6861 train_time:136142ms step_avg:168.91ms
step:817/1530 train_loss:3.7702 train_time:136314ms step_avg:168.91ms
step:818/1530 train_loss:3.5298 train_time:136487ms step_avg:168.92ms
step:819/1530 train_loss:3.7193 train_time:136663ms step_avg:168.93ms
step:820/1530 train_loss:3.4958 train_time:136840ms step_avg:168.94ms
step:821/1530 train_loss:3.5623 train_time:137012ms step_avg:168.94ms
step:822/1530 train_loss:3.6991 train_time:137188ms step_avg:168.95ms
step:823/1530 train_loss:3.5760 train_time:137364ms step_avg:168.96ms
step:824/1530 train_loss:3.5138 train_time:137537ms step_avg:168.96ms
step:825/1530 train_loss:3.6194 train_time:137712ms step_avg:168.97ms
step:826/1530 train_loss:3.4773 train_time:137889ms step_avg:168.98ms
step:827/1530 train_loss:3.7353 train_time:138064ms step_avg:168.99ms
step:828/1530 train_loss:3.6220 train_time:138238ms step_avg:169.00ms
step:829/1530 train_loss:3.6285 train_time:138413ms step_avg:169.00ms
step:830/1530 train_loss:3.5343 train_time:138588ms step_avg:169.01ms
step:831/1530 train_loss:3.6011 train_time:138762ms step_avg:169.02ms
step:832/1530 train_loss:3.5203 train_time:138936ms step_avg:169.02ms
step:833/1530 train_loss:3.6489 train_time:139110ms step_avg:169.03ms
step:834/1530 train_loss:3.4764 train_time:139284ms step_avg:169.03ms
step:835/1530 train_loss:3.4577 train_time:139459ms step_avg:169.04ms
step:836/1530 train_loss:3.7129 train_time:139635ms step_avg:169.05ms
step:837/1530 train_loss:3.3986 train_time:139808ms step_avg:169.05ms
step:838/1530 train_loss:3.5935 train_time:139983ms step_avg:169.06ms
step:839/1530 train_loss:3.4211 train_time:140156ms step_avg:169.07ms
step:840/1530 train_loss:3.4726 train_time:140328ms step_avg:169.07ms
step:841/1530 train_loss:3.5690 train_time:140502ms step_avg:169.08ms
step:842/1530 train_loss:3.5837 train_time:140676ms step_avg:169.08ms
step:843/1530 train_loss:3.5657 train_time:140848ms step_avg:169.09ms
step:844/1530 train_loss:3.4327 train_time:141020ms step_avg:169.09ms
step:845/1530 train_loss:3.6666 train_time:141195ms step_avg:169.10ms
step:846/1530 train_loss:3.5167 train_time:141369ms step_avg:169.10ms
step:847/1530 train_loss:3.4945 train_time:141544ms step_avg:169.11ms
step:848/1530 train_loss:3.6445 train_time:141716ms step_avg:169.11ms
step:849/1530 train_loss:3.4896 train_time:141890ms step_avg:169.12ms
step:850/1530 train_loss:3.4462 train_time:142064ms step_avg:169.12ms
step:851/1530 train_loss:3.7330 train_time:142238ms step_avg:169.13ms
step:852/1530 train_loss:3.4429 train_time:142410ms step_avg:169.13ms
step:853/1530 train_loss:3.5663 train_time:142583ms step_avg:169.14ms
step:854/1530 train_loss:3.6528 train_time:142757ms step_avg:169.14ms
step:855/1530 train_loss:3.5143 train_time:142932ms step_avg:169.15ms
step:856/1530 train_loss:3.5497 train_time:143106ms step_avg:169.16ms
step:857/1530 train_loss:3.6071 train_time:143281ms step_avg:169.16ms
step:858/1530 train_loss:3.4723 train_time:143456ms step_avg:169.17ms
step:859/1530 train_loss:3.5805 train_time:143630ms step_avg:169.18ms
step:860/1530 train_loss:3.5850 train_time:143803ms step_avg:169.18ms
step:861/1530 train_loss:3.6370 train_time:143979ms step_avg:169.19ms
step:862/1530 train_loss:3.6111 train_time:144155ms step_avg:169.20ms
step:863/1530 train_loss:3.5711 train_time:144329ms step_avg:169.20ms
step:864/1530 train_loss:3.3830 train_time:144504ms step_avg:169.21ms
step:865/1530 train_loss:3.6000 train_time:144676ms step_avg:169.21ms
step:866/1530 train_loss:3.8922 train_time:144852ms step_avg:169.22ms
step:867/1530 train_loss:3.4556 train_time:145025ms step_avg:169.22ms
step:868/1530 train_loss:3.6437 train_time:145198ms step_avg:169.23ms
step:869/1530 train_loss:3.6175 train_time:145370ms step_avg:169.23ms
step:870/1530 train_loss:3.4515 train_time:145544ms step_avg:169.24ms
step:871/1530 train_loss:3.3930 train_time:145719ms step_avg:169.24ms
step:872/1530 train_loss:3.6548 train_time:145894ms step_avg:169.25ms
step:873/1530 train_loss:3.4605 train_time:146066ms step_avg:169.25ms
step:874/1530 train_loss:3.2248 train_time:146244ms step_avg:169.26ms
step:875/1530 train_loss:3.6333 train_time:146417ms step_avg:169.27ms
step:875/1530 val_loss:3.5219 train_time:146467ms step_avg:169.33ms
step:876/1530 train_loss:3.4403 train_time:146590ms step_avg:169.27ms
step:877/1530 train_loss:3.6199 train_time:146765ms step_avg:169.28ms
step:878/1530 train_loss:3.4705 train_time:146940ms step_avg:169.29ms
step:879/1530 train_loss:3.6522 train_time:147113ms step_avg:169.29ms
step:880/1530 train_loss:3.3083 train_time:147285ms step_avg:169.29ms
step:881/1530 train_loss:3.4766 train_time:147458ms step_avg:169.30ms
step:882/1530 train_loss:3.6965 train_time:147631ms step_avg:169.30ms
step:883/1530 train_loss:3.8403 train_time:147804ms step_avg:169.31ms
step:884/1530 train_loss:3.5642 train_time:147979ms step_avg:169.31ms
step:885/1530 train_loss:3.4975 train_time:148152ms step_avg:169.32ms
step:886/1530 train_loss:3.5745 train_time:148325ms step_avg:169.32ms
step:887/1530 train_loss:4.0900 train_time:148499ms step_avg:169.33ms
step:888/1530 train_loss:3.8388 train_time:148679ms step_avg:169.34ms
step:889/1530 train_loss:3.5162 train_time:148853ms step_avg:169.34ms
step:890/1530 train_loss:3.5342 train_time:149026ms step_avg:169.35ms
step:891/1530 train_loss:3.3617 train_time:149201ms step_avg:169.35ms
step:892/1530 train_loss:3.7172 train_time:149373ms step_avg:169.36ms
step:893/1530 train_loss:3.4223 train_time:149545ms step_avg:169.36ms
step:894/1530 train_loss:3.6409 train_time:149724ms step_avg:169.37ms
step:895/1530 train_loss:3.6784 train_time:149897ms step_avg:169.38ms
step:896/1530 train_loss:3.4976 train_time:150069ms step_avg:169.38ms
step:897/1530 train_loss:3.5435 train_time:150245ms step_avg:169.39ms
step:898/1530 train_loss:3.5938 train_time:150423ms step_avg:169.39ms
step:899/1530 train_loss:3.4808 train_time:150596ms step_avg:169.40ms
step:900/1530 train_loss:3.4268 train_time:150767ms step_avg:169.40ms
step:901/1530 train_loss:3.6186 train_time:150941ms step_avg:169.41ms
step:902/1530 train_loss:3.6344 train_time:151115ms step_avg:169.41ms
step:903/1530 train_loss:3.5451 train_time:151290ms step_avg:169.42ms
step:904/1530 train_loss:3.4941 train_time:151464ms step_avg:169.42ms
step:905/1530 train_loss:3.5028 train_time:151634ms step_avg:169.42ms
step:906/1530 train_loss:3.7026 train_time:151809ms step_avg:169.43ms
step:907/1530 train_loss:3.5177 train_time:151982ms step_avg:169.43ms
step:908/1530 train_loss:3.5633 train_time:152155ms step_avg:169.44ms
step:909/1530 train_loss:3.4561 train_time:152331ms step_avg:169.44ms
step:910/1530 train_loss:3.5310 train_time:152512ms step_avg:169.46ms
step:911/1530 train_loss:3.6478 train_time:152687ms step_avg:169.46ms
step:912/1530 train_loss:3.5925 train_time:152866ms step_avg:169.47ms
step:913/1530 train_loss:3.4592 train_time:153045ms step_avg:169.48ms
step:914/1530 train_loss:3.7452 train_time:153224ms step_avg:169.50ms
step:915/1530 train_loss:3.5398 train_time:153402ms step_avg:169.51ms
step:916/1530 train_loss:3.6199 train_time:153578ms step_avg:169.51ms
step:917/1530 train_loss:3.6011 train_time:153752ms step_avg:169.52ms
step:918/1530 train_loss:4.8177 train_time:153931ms step_avg:169.53ms
step:919/1530 train_loss:3.4992 train_time:154110ms step_avg:169.54ms
step:920/1530 train_loss:3.5909 train_time:154285ms step_avg:169.54ms
step:921/1530 train_loss:3.5577 train_time:154464ms step_avg:169.55ms
step:922/1530 train_loss:3.5832 train_time:154641ms step_avg:169.56ms
step:923/1530 train_loss:3.6169 train_time:154814ms step_avg:169.57ms
step:924/1530 train_loss:3.6817 train_time:154989ms step_avg:169.57ms
step:925/1530 train_loss:3.6486 train_time:155163ms step_avg:169.58ms
step:926/1530 train_loss:3.5564 train_time:155335ms step_avg:169.58ms
step:927/1530 train_loss:3.5555 train_time:155511ms step_avg:169.59ms
step:928/1530 train_loss:3.7827 train_time:155687ms step_avg:169.59ms
step:929/1530 train_loss:3.6097 train_time:155861ms step_avg:169.60ms
step:930/1530 train_loss:3.4026 train_time:156036ms step_avg:169.60ms
step:931/1530 train_loss:3.4971 train_time:156210ms step_avg:169.61ms
step:932/1530 train_loss:3.6516 train_time:156388ms step_avg:169.62ms
step:933/1530 train_loss:3.3639 train_time:156563ms step_avg:169.62ms
step:934/1530 train_loss:3.5864 train_time:156741ms step_avg:169.63ms
step:935/1530 train_loss:3.4420 train_time:156920ms step_avg:169.64ms
step:936/1530 train_loss:3.5205 train_time:157096ms step_avg:169.65ms
step:937/1530 train_loss:3.6248 train_time:157275ms step_avg:169.66ms
step:938/1530 train_loss:3.5412 train_time:157448ms step_avg:169.66ms
step:939/1530 train_loss:3.6821 train_time:157629ms step_avg:169.68ms
step:940/1530 train_loss:3.4830 train_time:157805ms step_avg:169.68ms
step:941/1530 train_loss:3.5506 train_time:157979ms step_avg:169.69ms
step:942/1530 train_loss:3.3551 train_time:158154ms step_avg:169.69ms
step:943/1530 train_loss:3.7140 train_time:158334ms step_avg:169.70ms
step:944/1530 train_loss:3.4053 train_time:158647ms step_avg:169.86ms
step:945/1530 train_loss:3.4303 train_time:158831ms step_avg:169.87ms
step:946/1530 train_loss:5.0684 train_time:159012ms step_avg:169.89ms
step:947/1530 train_loss:3.6009 train_time:159188ms step_avg:169.89ms
step:948/1530 train_loss:3.4898 train_time:159364ms step_avg:169.90ms
step:949/1530 train_loss:3.3759 train_time:159690ms step_avg:170.06ms
step:950/1530 train_loss:3.4425 train_time:159865ms step_avg:170.07ms
step:951/1530 train_loss:3.4121 train_time:160045ms step_avg:170.08ms
step:952/1530 train_loss:3.4775 train_time:160223ms step_avg:170.09ms
step:953/1530 train_loss:3.5693 train_time:160399ms step_avg:170.09ms
step:954/1530 train_loss:3.4441 train_time:160579ms step_avg:170.10ms
step:955/1530 train_loss:3.4788 train_time:160752ms step_avg:170.11ms
step:956/1530 train_loss:3.4437 train_time:160927ms step_avg:170.11ms
step:957/1530 train_loss:3.4954 train_time:161108ms step_avg:170.12ms
step:958/1530 train_loss:3.5044 train_time:161289ms step_avg:170.14ms
step:959/1530 train_loss:3.5137 train_time:161465ms step_avg:170.14ms
step:960/1530 train_loss:3.4072 train_time:161644ms step_avg:170.15ms
step:961/1530 train_loss:3.6475 train_time:161818ms step_avg:170.16ms
step:962/1530 train_loss:3.5920 train_time:161992ms step_avg:170.16ms
step:963/1530 train_loss:3.5676 train_time:162169ms step_avg:170.17ms
step:964/1530 train_loss:3.4267 train_time:162347ms step_avg:170.18ms
step:965/1530 train_loss:3.4753 train_time:162521ms step_avg:170.18ms
step:966/1530 train_loss:3.7073 train_time:162694ms step_avg:170.18ms
step:967/1530 train_loss:3.5253 train_time:162868ms step_avg:170.19ms
step:968/1530 train_loss:3.5163 train_time:163044ms step_avg:170.19ms
step:969/1530 train_loss:3.5833 train_time:163220ms step_avg:170.20ms
step:970/1530 train_loss:3.3810 train_time:163391ms step_avg:170.20ms
step:971/1530 train_loss:3.5310 train_time:163566ms step_avg:170.20ms
step:972/1530 train_loss:3.4794 train_time:163739ms step_avg:170.21ms
step:973/1530 train_loss:3.5405 train_time:163912ms step_avg:170.21ms
step:974/1530 train_loss:3.5902 train_time:164088ms step_avg:170.22ms
step:975/1530 train_loss:3.4642 train_time:164264ms step_avg:170.22ms
step:976/1530 train_loss:3.6714 train_time:164438ms step_avg:170.23ms
step:977/1530 train_loss:3.5741 train_time:164612ms step_avg:170.23ms
step:978/1530 train_loss:3.3635 train_time:164788ms step_avg:170.24ms
step:979/1530 train_loss:3.6323 train_time:164964ms step_avg:170.24ms
step:980/1530 train_loss:3.4237 train_time:165139ms step_avg:170.25ms
step:981/1530 train_loss:3.5820 train_time:165316ms step_avg:170.25ms
step:982/1530 train_loss:3.5452 train_time:165490ms step_avg:170.26ms
step:983/1530 train_loss:3.5191 train_time:165666ms step_avg:170.26ms
step:984/1530 train_loss:3.4946 train_time:165843ms step_avg:170.27ms
step:985/1530 train_loss:3.5801 train_time:166022ms step_avg:170.28ms
step:986/1530 train_loss:3.4132 train_time:166196ms step_avg:170.28ms
step:987/1530 train_loss:3.4838 train_time:166368ms step_avg:170.29ms
step:988/1530 train_loss:3.4789 train_time:166543ms step_avg:170.29ms
step:989/1530 train_loss:3.4193 train_time:166718ms step_avg:170.29ms
step:990/1530 train_loss:3.6613 train_time:166892ms step_avg:170.30ms
step:991/1530 train_loss:3.4673 train_time:167065ms step_avg:170.30ms
step:992/1530 train_loss:3.4486 train_time:167246ms step_avg:170.31ms
step:993/1530 train_loss:3.5010 train_time:167425ms step_avg:170.32ms
step:994/1530 train_loss:3.5987 train_time:167598ms step_avg:170.32ms
step:995/1530 train_loss:3.5265 train_time:167770ms step_avg:170.32ms
step:996/1530 train_loss:3.4550 train_time:167944ms step_avg:170.33ms
step:997/1530 train_loss:3.7565 train_time:168121ms step_avg:170.34ms
step:998/1530 train_loss:3.4394 train_time:168292ms step_avg:170.34ms
step:999/1530 train_loss:3.5864 train_time:168467ms step_avg:170.34ms
step:1000/1530 train_loss:3.4416 train_time:168645ms step_avg:170.35ms
step:1000/1530 val_loss:3.4675 train_time:168697ms step_avg:170.40ms
step:1001/1530 train_loss:3.5022 train_time:168822ms step_avg:170.36ms
step:1002/1530 train_loss:3.3741 train_time:168996ms step_avg:170.36ms
step:1003/1530 train_loss:3.5553 train_time:169172ms step_avg:170.36ms
step:1004/1530 train_loss:3.6039 train_time:169348ms step_avg:170.37ms
step:1005/1530 train_loss:3.3900 train_time:169523ms step_avg:170.37ms
step:1006/1530 train_loss:3.4651 train_time:169699ms step_avg:170.38ms
step:1007/1530 train_loss:3.4439 train_time:169874ms step_avg:170.39ms
step:1008/1530 train_loss:3.5627 train_time:170050ms step_avg:170.39ms
step:1009/1530 train_loss:3.6629 train_time:170228ms step_avg:170.40ms
step:1010/1530 train_loss:3.5647 train_time:170401ms step_avg:170.40ms
step:1011/1530 train_loss:3.5352 train_time:170573ms step_avg:170.40ms
step:1012/1530 train_loss:3.3936 train_time:170748ms step_avg:170.41ms
step:1013/1530 train_loss:3.5335 train_time:170923ms step_avg:170.41ms
step:1014/1530 train_loss:3.6275 train_time:171100ms step_avg:170.42ms
step:1015/1530 train_loss:3.3268 train_time:171277ms step_avg:170.42ms
step:1016/1530 train_loss:3.4167 train_time:171451ms step_avg:170.43ms
step:1017/1530 train_loss:3.3999 train_time:171629ms step_avg:170.44ms
step:1018/1530 train_loss:3.3972 train_time:171804ms step_avg:170.44ms
step:1019/1530 train_loss:3.5208 train_time:171979ms step_avg:170.45ms
step:1020/1530 train_loss:3.3791 train_time:172157ms step_avg:170.45ms
step:1021/1530 train_loss:3.3552 train_time:172334ms step_avg:170.46ms
step:1022/1530 train_loss:3.4789 train_time:172510ms step_avg:170.46ms
step:1023/1530 train_loss:3.5102 train_time:172686ms step_avg:170.47ms
step:1024/1530 train_loss:3.4733 train_time:172862ms step_avg:170.48ms
step:1025/1530 train_loss:3.4820 train_time:173041ms step_avg:170.48ms
step:1026/1530 train_loss:3.6191 train_time:173217ms step_avg:170.49ms
step:1027/1530 train_loss:3.3267 train_time:173393ms step_avg:170.49ms
step:1028/1530 train_loss:3.3988 train_time:173576ms step_avg:170.51ms
step:1029/1530 train_loss:3.3166 train_time:173757ms step_avg:170.52ms
step:1030/1530 train_loss:3.5381 train_time:173935ms step_avg:170.52ms
step:1031/1530 train_loss:3.5075 train_time:174109ms step_avg:170.53ms
step:1032/1530 train_loss:3.6875 train_time:174291ms step_avg:170.54ms
step:1033/1530 train_loss:3.4913 train_time:174467ms step_avg:170.54ms
step:1034/1530 train_loss:3.4004 train_time:174644ms step_avg:170.55ms
step:1035/1530 train_loss:3.4470 train_time:174823ms step_avg:170.56ms
step:1036/1530 train_loss:3.4823 train_time:174999ms step_avg:170.56ms
step:1037/1530 train_loss:3.7890 train_time:175178ms step_avg:170.57ms
step:1038/1530 train_loss:3.6201 train_time:175357ms step_avg:170.58ms
step:1039/1530 train_loss:3.5110 train_time:175539ms step_avg:170.59ms
step:1040/1530 train_loss:3.4107 train_time:175716ms step_avg:170.60ms
step:1041/1530 train_loss:3.4894 train_time:175895ms step_avg:170.61ms
step:1042/1530 train_loss:3.5214 train_time:176070ms step_avg:170.61ms
step:1043/1530 train_loss:3.4506 train_time:176245ms step_avg:170.61ms
step:1044/1530 train_loss:3.4548 train_time:176421ms step_avg:170.62ms
step:1045/1530 train_loss:3.5138 train_time:176599ms step_avg:170.63ms
step:1046/1530 train_loss:3.4284 train_time:176774ms step_avg:170.63ms
step:1047/1530 train_loss:3.6347 train_time:176950ms step_avg:170.64ms
step:1048/1530 train_loss:3.4969 train_time:177126ms step_avg:170.64ms
step:1049/1530 train_loss:3.4060 train_time:177302ms step_avg:170.65ms
step:1050/1530 train_loss:3.3918 train_time:177479ms step_avg:170.65ms
step:1051/1530 train_loss:3.4955 train_time:177657ms step_avg:170.66ms
step:1052/1530 train_loss:3.3631 train_time:177836ms step_avg:170.67ms
step:1053/1530 train_loss:3.6879 train_time:178013ms step_avg:170.67ms
step:1054/1530 train_loss:3.5415 train_time:178190ms step_avg:170.68ms
step:1055/1530 train_loss:3.3898 train_time:178365ms step_avg:170.68ms
step:1056/1530 train_loss:3.4950 train_time:178540ms step_avg:170.69ms
step:1057/1530 train_loss:3.5817 train_time:178717ms step_avg:170.69ms
step:1058/1530 train_loss:3.3043 train_time:178895ms step_avg:170.70ms
step:1059/1530 train_loss:3.3677 train_time:179077ms step_avg:170.71ms
step:1060/1530 train_loss:3.4334 train_time:179254ms step_avg:170.72ms
step:1061/1530 train_loss:3.4184 train_time:179428ms step_avg:170.72ms
step:1062/1530 train_loss:3.3800 train_time:179603ms step_avg:170.73ms
step:1063/1530 train_loss:3.4565 train_time:179777ms step_avg:170.73ms
step:1064/1530 train_loss:3.3800 train_time:179953ms step_avg:170.73ms
step:1065/1530 train_loss:3.3660 train_time:180132ms step_avg:170.74ms
step:1066/1530 train_loss:3.4110 train_time:180308ms step_avg:170.75ms
step:1067/1530 train_loss:3.2853 train_time:180487ms step_avg:170.75ms
step:1068/1530 train_loss:3.4365 train_time:180662ms step_avg:170.76ms
step:1069/1530 train_loss:3.3009 train_time:180843ms step_avg:170.77ms
step:1070/1530 train_loss:3.5691 train_time:181019ms step_avg:170.77ms
step:1071/1530 train_loss:3.5133 train_time:181197ms step_avg:170.78ms
step:1072/1530 train_loss:3.4391 train_time:181372ms step_avg:170.78ms
step:1073/1530 train_loss:3.5219 train_time:181544ms step_avg:170.78ms
step:1074/1530 train_loss:3.4287 train_time:181721ms step_avg:170.79ms
step:1075/1530 train_loss:3.3969 train_time:181898ms step_avg:170.80ms
step:1076/1530 train_loss:3.7964 train_time:182075ms step_avg:170.80ms
step:1077/1530 train_loss:3.4313 train_time:182251ms step_avg:170.81ms
step:1078/1530 train_loss:3.0885 train_time:182436ms step_avg:170.82ms
step:1079/1530 train_loss:3.5297 train_time:182614ms step_avg:170.83ms
step:1080/1530 train_loss:3.4266 train_time:182792ms step_avg:170.83ms
step:1081/1530 train_loss:3.5020 train_time:182965ms step_avg:170.84ms
step:1082/1530 train_loss:3.5923 train_time:183141ms step_avg:170.84ms
step:1083/1530 train_loss:3.4936 train_time:183317ms step_avg:170.84ms
step:1084/1530 train_loss:3.4675 train_time:183494ms step_avg:170.85ms
step:1085/1530 train_loss:3.4286 train_time:183667ms step_avg:170.85ms
step:1086/1530 train_loss:3.6258 train_time:183844ms step_avg:170.86ms
step:1087/1530 train_loss:3.5017 train_time:184019ms step_avg:170.86ms
step:1088/1530 train_loss:3.3692 train_time:184197ms step_avg:170.87ms
step:1089/1530 train_loss:3.3763 train_time:184376ms step_avg:170.88ms
step:1090/1530 train_loss:3.4801 train_time:184556ms step_avg:170.88ms
step:1091/1530 train_loss:3.2834 train_time:184732ms step_avg:170.89ms
step:1092/1530 train_loss:3.4829 train_time:184908ms step_avg:170.89ms
step:1093/1530 train_loss:3.6015 train_time:185083ms step_avg:170.90ms
step:1094/1530 train_loss:3.4502 train_time:185258ms step_avg:170.90ms
step:1095/1530 train_loss:3.4181 train_time:185434ms step_avg:170.91ms
step:1096/1530 train_loss:3.4252 train_time:185610ms step_avg:170.91ms
step:1097/1530 train_loss:3.4898 train_time:185786ms step_avg:170.92ms
step:1098/1530 train_loss:3.5642 train_time:185964ms step_avg:170.92ms
step:1099/1530 train_loss:3.5258 train_time:186142ms step_avg:170.93ms
step:1100/1530 train_loss:3.4247 train_time:186322ms step_avg:170.94ms
step:1101/1530 train_loss:3.2904 train_time:186500ms step_avg:170.94ms
step:1102/1530 train_loss:3.3051 train_time:186680ms step_avg:170.95ms
step:1103/1530 train_loss:3.4419 train_time:186860ms step_avg:170.96ms
step:1104/1530 train_loss:3.3201 train_time:187037ms step_avg:170.97ms
step:1105/1530 train_loss:4.0560 train_time:187216ms step_avg:170.97ms
step:1106/1530 train_loss:3.2265 train_time:187393ms step_avg:170.98ms
step:1107/1530 train_loss:3.5676 train_time:187569ms step_avg:170.98ms
step:1108/1530 train_loss:3.3487 train_time:187742ms step_avg:170.99ms
step:1109/1530 train_loss:3.5005 train_time:187918ms step_avg:170.99ms
step:1110/1530 train_loss:3.4285 train_time:188092ms step_avg:170.99ms
step:1111/1530 train_loss:3.4812 train_time:188265ms step_avg:170.99ms
step:1112/1530 train_loss:3.5544 train_time:188443ms step_avg:171.00ms
step:1113/1530 train_loss:3.4324 train_time:188625ms step_avg:171.01ms
step:1114/1530 train_loss:3.3665 train_time:188804ms step_avg:171.02ms
step:1115/1530 train_loss:3.2368 train_time:188981ms step_avg:171.02ms
step:1116/1530 train_loss:3.4266 train_time:189154ms step_avg:171.03ms
step:1117/1530 train_loss:3.5933 train_time:189334ms step_avg:171.03ms
step:1118/1530 train_loss:3.6205 train_time:189511ms step_avg:171.04ms
step:1119/1530 train_loss:3.4797 train_time:189683ms step_avg:171.04ms
step:1120/1530 train_loss:3.4899 train_time:189862ms step_avg:171.05ms
step:1121/1530 train_loss:3.3913 train_time:190041ms step_avg:171.05ms
step:1122/1530 train_loss:3.4556 train_time:190217ms step_avg:171.06ms
step:1123/1530 train_loss:3.5790 train_time:190394ms step_avg:171.06ms
step:1124/1530 train_loss:3.3373 train_time:190569ms step_avg:171.07ms
step:1125/1530 train_loss:3.2334 train_time:190745ms step_avg:171.07ms
step:1125/1530 val_loss:3.4097 train_time:190795ms step_avg:171.12ms
step:1126/1530 train_loss:3.4794 train_time:190923ms step_avg:171.08ms
step:1127/1530 train_loss:3.6767 train_time:191102ms step_avg:171.09ms
step:1128/1530 train_loss:3.2324 train_time:191280ms step_avg:171.09ms
step:1129/1530 train_loss:3.5566 train_time:191460ms step_avg:171.10ms
step:1130/1530 train_loss:3.3772 train_time:191637ms step_avg:171.10ms
step:1131/1530 train_loss:3.4012 train_time:191819ms step_avg:171.11ms
step:1132/1530 train_loss:3.3652 train_time:191993ms step_avg:171.12ms
step:1133/1530 train_loss:3.4858 train_time:192302ms step_avg:171.24ms
step:1134/1530 train_loss:3.4516 train_time:192487ms step_avg:171.25ms
step:1135/1530 train_loss:3.5194 train_time:192665ms step_avg:171.26ms
step:1136/1530 train_loss:3.5657 train_time:192844ms step_avg:171.26ms
step:1137/1530 train_loss:3.4560 train_time:193021ms step_avg:171.27ms
step:1138/1530 train_loss:3.3560 train_time:193201ms step_avg:171.28ms
step:1139/1530 train_loss:3.6545 train_time:193526ms step_avg:171.41ms
step:1140/1530 train_loss:3.4534 train_time:193705ms step_avg:171.42ms
step:1141/1530 train_loss:3.5938 train_time:193884ms step_avg:171.43ms
step:1142/1530 train_loss:3.4393 train_time:194062ms step_avg:171.43ms
step:1143/1530 train_loss:3.3656 train_time:194241ms step_avg:171.44ms
step:1144/1530 train_loss:3.4469 train_time:194416ms step_avg:171.44ms
step:1145/1530 train_loss:3.5928 train_time:194589ms step_avg:171.44ms
step:1146/1530 train_loss:3.5608 train_time:194770ms step_avg:171.45ms
step:1147/1530 train_loss:3.4916 train_time:194949ms step_avg:171.46ms
step:1148/1530 train_loss:3.5005 train_time:195126ms step_avg:171.46ms
step:1149/1530 train_loss:3.3261 train_time:195308ms step_avg:171.47ms
step:1150/1530 train_loss:3.3736 train_time:195483ms step_avg:171.48ms
step:1151/1530 train_loss:3.3173 train_time:195662ms step_avg:171.48ms
step:1152/1530 train_loss:3.4019 train_time:195845ms step_avg:171.49ms
step:1153/1530 train_loss:3.4321 train_time:196026ms step_avg:171.50ms
step:1154/1530 train_loss:3.5215 train_time:196204ms step_avg:171.51ms
step:1155/1530 train_loss:3.3209 train_time:196385ms step_avg:171.52ms
step:1156/1530 train_loss:3.5338 train_time:196567ms step_avg:171.52ms
step:1157/1530 train_loss:3.4932 train_time:196746ms step_avg:171.53ms
step:1158/1530 train_loss:3.2543 train_time:196923ms step_avg:171.54ms
step:1159/1530 train_loss:3.3515 train_time:197100ms step_avg:171.54ms
step:1160/1530 train_loss:3.3403 train_time:197274ms step_avg:171.54ms
step:1161/1530 train_loss:3.0927 train_time:197454ms step_avg:171.55ms
step:1162/1530 train_loss:3.4196 train_time:197631ms step_avg:171.55ms
step:1163/1530 train_loss:3.3926 train_time:197808ms step_avg:171.56ms
step:1164/1530 train_loss:3.2918 train_time:197985ms step_avg:171.56ms
step:1165/1530 train_loss:3.2464 train_time:198163ms step_avg:171.57ms
step:1166/1530 train_loss:3.3894 train_time:198344ms step_avg:171.58ms
step:1167/1530 train_loss:3.4081 train_time:198520ms step_avg:171.58ms
step:1168/1530 train_loss:3.7206 train_time:198695ms step_avg:171.59ms
step:1169/1530 train_loss:3.3794 train_time:198872ms step_avg:171.59ms
step:1170/1530 train_loss:3.3914 train_time:199049ms step_avg:171.59ms
step:1171/1530 train_loss:3.2913 train_time:199226ms step_avg:171.60ms
step:1172/1530 train_loss:3.4260 train_time:199401ms step_avg:171.60ms
step:1173/1530 train_loss:3.5379 train_time:199582ms step_avg:171.61ms
step:1174/1530 train_loss:3.3861 train_time:199767ms step_avg:171.62ms
step:1175/1530 train_loss:3.3655 train_time:199946ms step_avg:171.63ms
step:1176/1530 train_loss:3.4286 train_time:200127ms step_avg:171.64ms
step:1177/1530 train_loss:3.4505 train_time:200309ms step_avg:171.64ms
step:1178/1530 train_loss:3.4973 train_time:200484ms step_avg:171.65ms
step:1179/1530 train_loss:3.4001 train_time:200659ms step_avg:171.65ms
step:1180/1530 train_loss:3.3576 train_time:200847ms step_avg:171.66ms
step:1181/1530 train_loss:3.3403 train_time:201025ms step_avg:171.67ms
step:1182/1530 train_loss:3.3810 train_time:201204ms step_avg:171.68ms
step:1183/1530 train_loss:3.3343 train_time:201382ms step_avg:171.68ms
step:1184/1530 train_loss:3.5096 train_time:201558ms step_avg:171.68ms
step:1185/1530 train_loss:3.5475 train_time:201740ms step_avg:171.69ms
step:1186/1530 train_loss:3.3656 train_time:201921ms step_avg:171.70ms
step:1187/1530 train_loss:3.4176 train_time:202107ms step_avg:171.71ms
step:1188/1530 train_loss:3.4435 train_time:202284ms step_avg:171.72ms
step:1189/1530 train_loss:3.2796 train_time:202465ms step_avg:171.73ms
step:1190/1530 train_loss:3.4425 train_time:202645ms step_avg:171.73ms
step:1191/1530 train_loss:3.5818 train_time:202826ms step_avg:171.74ms
step:1192/1530 train_loss:3.3922 train_time:203002ms step_avg:171.74ms
step:1193/1530 train_loss:3.2736 train_time:203178ms step_avg:171.75ms
step:1194/1530 train_loss:3.5568 train_time:203355ms step_avg:171.75ms
step:1195/1530 train_loss:3.3708 train_time:203536ms step_avg:171.76ms
step:1196/1530 train_loss:3.3836 train_time:203724ms step_avg:171.77ms
step:1197/1530 train_loss:3.2888 train_time:203905ms step_avg:171.78ms
step:1198/1530 train_loss:3.3018 train_time:204090ms step_avg:171.79ms
step:1199/1530 train_loss:3.3466 train_time:204270ms step_avg:171.80ms
step:1200/1530 train_loss:3.4423 train_time:204447ms step_avg:171.80ms
step:1201/1530 train_loss:3.4790 train_time:204626ms step_avg:171.81ms
step:1202/1530 train_loss:3.6109 train_time:204815ms step_avg:171.82ms
step:1203/1530 train_loss:3.4026 train_time:204995ms step_avg:171.83ms
step:1204/1530 train_loss:3.3081 train_time:205176ms step_avg:171.84ms
step:1205/1530 train_loss:3.4379 train_time:205352ms step_avg:171.84ms
step:1206/1530 train_loss:3.4703 train_time:205529ms step_avg:171.85ms
step:1207/1530 train_loss:3.5172 train_time:205707ms step_avg:171.85ms
step:1208/1530 train_loss:3.3971 train_time:205883ms step_avg:171.86ms
step:1209/1530 train_loss:3.2455 train_time:206063ms step_avg:171.86ms
step:1210/1530 train_loss:3.3067 train_time:206243ms step_avg:171.87ms
step:1211/1530 train_loss:3.3915 train_time:206420ms step_avg:171.87ms
step:1212/1530 train_loss:3.3952 train_time:206594ms step_avg:171.88ms
step:1213/1530 train_loss:3.4151 train_time:206773ms step_avg:171.88ms
step:1214/1530 train_loss:3.2638 train_time:206953ms step_avg:171.89ms
step:1215/1530 train_loss:3.3966 train_time:207129ms step_avg:171.89ms
step:1216/1530 train_loss:3.3316 train_time:207308ms step_avg:171.90ms
step:1217/1530 train_loss:3.3219 train_time:207484ms step_avg:171.90ms
step:1218/1530 train_loss:3.4064 train_time:207664ms step_avg:171.91ms
step:1219/1530 train_loss:3.2547 train_time:207848ms step_avg:171.92ms
step:1220/1530 train_loss:3.4800 train_time:208024ms step_avg:171.92ms
step:1221/1530 train_loss:3.5026 train_time:208200ms step_avg:171.92ms
step:1222/1530 train_loss:3.4409 train_time:208375ms step_avg:171.93ms
step:1223/1530 train_loss:3.2968 train_time:208552ms step_avg:171.93ms
step:1224/1530 train_loss:3.2547 train_time:208735ms step_avg:171.94ms
step:1225/1530 train_loss:3.3686 train_time:208911ms step_avg:171.94ms
step:1226/1530 train_loss:3.3360 train_time:209090ms step_avg:171.95ms
step:1227/1530 train_loss:3.2771 train_time:209269ms step_avg:171.95ms
step:1228/1530 train_loss:3.4429 train_time:209445ms step_avg:171.96ms
step:1229/1530 train_loss:3.3659 train_time:209626ms step_avg:171.97ms
step:1230/1530 train_loss:3.3959 train_time:209808ms step_avg:171.97ms
step:1231/1530 train_loss:3.5758 train_time:209989ms step_avg:171.98ms
step:1232/1530 train_loss:3.4992 train_time:210170ms step_avg:171.99ms
step:1233/1530 train_loss:3.4308 train_time:210347ms step_avg:171.99ms
step:1234/1530 train_loss:3.5851 train_time:210526ms step_avg:172.00ms
step:1235/1530 train_loss:3.3266 train_time:210707ms step_avg:172.01ms
step:1236/1530 train_loss:3.2926 train_time:210884ms step_avg:172.01ms
step:1237/1530 train_loss:3.2737 train_time:211061ms step_avg:172.01ms
step:1238/1530 train_loss:3.2805 train_time:211246ms step_avg:172.02ms
step:1239/1530 train_loss:3.3317 train_time:211425ms step_avg:172.03ms
step:1240/1530 train_loss:3.3822 train_time:211602ms step_avg:172.03ms
step:1241/1530 train_loss:3.4230 train_time:211780ms step_avg:172.04ms
step:1242/1530 train_loss:3.2992 train_time:211956ms step_avg:172.04ms
step:1243/1530 train_loss:3.4078 train_time:212135ms step_avg:172.05ms
step:1244/1530 train_loss:3.4092 train_time:212308ms step_avg:172.05ms
step:1245/1530 train_loss:3.4167 train_time:212484ms step_avg:172.05ms
step:1246/1530 train_loss:3.2435 train_time:212664ms step_avg:172.06ms
step:1247/1530 train_loss:3.3730 train_time:212839ms step_avg:172.06ms
step:1248/1530 train_loss:3.4269 train_time:213014ms step_avg:172.06ms
step:1249/1530 train_loss:3.4257 train_time:213192ms step_avg:172.07ms
step:1250/1530 train_loss:3.3064 train_time:213371ms step_avg:172.07ms
step:1250/1530 val_loss:3.3562 train_time:213425ms step_avg:172.12ms
step:1251/1530 train_loss:3.4919 train_time:213558ms step_avg:172.09ms
step:1252/1530 train_loss:3.3600 train_time:213735ms step_avg:172.09ms
step:1253/1530 train_loss:3.3091 train_time:213911ms step_avg:172.09ms
step:1254/1530 train_loss:3.4172 train_time:214091ms step_avg:172.10ms
step:1255/1530 train_loss:3.5204 train_time:214281ms step_avg:172.11ms
step:1256/1530 train_loss:3.3014 train_time:214463ms step_avg:172.12ms
step:1257/1530 train_loss:3.3727 train_time:214641ms step_avg:172.13ms
step:1258/1530 train_loss:3.3654 train_time:214826ms step_avg:172.14ms
step:1259/1530 train_loss:3.3291 train_time:215003ms step_avg:172.14ms
step:1260/1530 train_loss:3.2152 train_time:215180ms step_avg:172.14ms
step:1261/1530 train_loss:3.3042 train_time:215361ms step_avg:172.15ms
step:1262/1530 train_loss:3.3295 train_time:215545ms step_avg:172.16ms
step:1263/1530 train_loss:3.2401 train_time:215728ms step_avg:172.17ms
step:1264/1530 train_loss:3.4440 train_time:215903ms step_avg:172.17ms
step:1265/1530 train_loss:3.4259 train_time:216079ms step_avg:172.17ms
step:1266/1530 train_loss:3.4428 train_time:216259ms step_avg:172.18ms
step:1267/1530 train_loss:3.3707 train_time:216440ms step_avg:172.19ms
step:1268/1530 train_loss:3.4120 train_time:216620ms step_avg:172.19ms
step:1269/1530 train_loss:3.2508 train_time:216805ms step_avg:172.20ms
step:1270/1530 train_loss:3.1082 train_time:216982ms step_avg:172.21ms
step:1271/1530 train_loss:3.4061 train_time:217161ms step_avg:172.21ms
step:1272/1530 train_loss:3.3533 train_time:217339ms step_avg:172.22ms
step:1273/1530 train_loss:3.3816 train_time:217521ms step_avg:172.23ms
step:1274/1530 train_loss:3.3609 train_time:217701ms step_avg:172.23ms
step:1275/1530 train_loss:3.4344 train_time:217878ms step_avg:172.24ms
step:1276/1530 train_loss:3.4722 train_time:218053ms step_avg:172.24ms
step:1277/1530 train_loss:3.4064 train_time:218234ms step_avg:172.24ms
step:1278/1530 train_loss:3.4094 train_time:218408ms step_avg:172.25ms
step:1279/1530 train_loss:3.2709 train_time:218589ms step_avg:172.25ms
step:1280/1530 train_loss:3.3697 train_time:218776ms step_avg:172.26ms
step:1281/1530 train_loss:3.4296 train_time:218952ms step_avg:172.27ms
step:1282/1530 train_loss:3.4687 train_time:219127ms step_avg:172.27ms
step:1283/1530 train_loss:3.3360 train_time:219306ms step_avg:172.27ms
step:1284/1530 train_loss:3.3712 train_time:219482ms step_avg:172.28ms
step:1285/1530 train_loss:3.3664 train_time:219662ms step_avg:172.28ms
step:1286/1530 train_loss:3.3361 train_time:219839ms step_avg:172.29ms
step:1287/1530 train_loss:3.4879 train_time:220017ms step_avg:172.29ms
step:1288/1530 train_loss:3.2993 train_time:220199ms step_avg:172.30ms
step:1289/1530 train_loss:3.3773 train_time:220384ms step_avg:172.31ms
step:1290/1530 train_loss:3.4578 train_time:220567ms step_avg:172.32ms
step:1291/1530 train_loss:3.3853 train_time:220746ms step_avg:172.32ms
step:1292/1530 train_loss:3.4767 train_time:220928ms step_avg:172.33ms
step:1293/1530 train_loss:3.5145 train_time:221108ms step_avg:172.34ms
step:1294/1530 train_loss:3.4630 train_time:221288ms step_avg:172.34ms
step:1295/1530 train_loss:3.2829 train_time:221467ms step_avg:172.35ms
step:1296/1530 train_loss:3.3741 train_time:221648ms step_avg:172.35ms
step:1297/1530 train_loss:3.2767 train_time:221828ms step_avg:172.36ms
step:1298/1530 train_loss:3.2744 train_time:222007ms step_avg:172.37ms
step:1299/1530 train_loss:3.3959 train_time:222184ms step_avg:172.37ms
step:1300/1530 train_loss:3.4039 train_time:222361ms step_avg:172.37ms
step:1301/1530 train_loss:3.4083 train_time:222539ms step_avg:172.38ms
step:1302/1530 train_loss:3.5768 train_time:222721ms step_avg:172.38ms
step:1303/1530 train_loss:3.3081 train_time:222905ms step_avg:172.39ms
step:1304/1530 train_loss:3.5122 train_time:223085ms step_avg:172.40ms
step:1305/1530 train_loss:3.2623 train_time:223261ms step_avg:172.40ms
step:1306/1530 train_loss:3.4559 train_time:223442ms step_avg:172.41ms
step:1307/1530 train_loss:3.4543 train_time:223618ms step_avg:172.41ms
step:1308/1530 train_loss:3.2889 train_time:223795ms step_avg:172.42ms
step:1309/1530 train_loss:3.3092 train_time:223975ms step_avg:172.42ms
step:1310/1530 train_loss:3.2921 train_time:224151ms step_avg:172.42ms
step:1311/1530 train_loss:3.2978 train_time:224330ms step_avg:172.43ms
step:1312/1530 train_loss:3.3747 train_time:224510ms step_avg:172.43ms
step:1313/1530 train_loss:3.3421 train_time:224686ms step_avg:172.44ms
step:1314/1530 train_loss:3.0475 train_time:224868ms step_avg:172.45ms
step:1315/1530 train_loss:3.2753 train_time:225045ms step_avg:172.45ms
step:1316/1530 train_loss:3.3988 train_time:225220ms step_avg:172.45ms
step:1317/1530 train_loss:3.4191 train_time:225400ms step_avg:172.46ms
step:1318/1530 train_loss:3.3016 train_time:225585ms step_avg:172.47ms
step:1319/1530 train_loss:3.4268 train_time:225765ms step_avg:172.47ms
step:1320/1530 train_loss:3.4618 train_time:225947ms step_avg:172.48ms
step:1321/1530 train_loss:3.3634 train_time:226125ms step_avg:172.48ms
step:1322/1530 train_loss:3.3257 train_time:226435ms step_avg:172.59ms
step:1323/1530 train_loss:3.3206 train_time:226624ms step_avg:172.60ms
step:1324/1530 train_loss:3.4411 train_time:226803ms step_avg:172.61ms
step:1325/1530 train_loss:3.4946 train_time:226986ms step_avg:172.61ms
step:1326/1530 train_loss:3.2136 train_time:227166ms step_avg:172.62ms
step:1327/1530 train_loss:3.1696 train_time:227342ms step_avg:172.62ms
step:1328/1530 train_loss:3.4943 train_time:227522ms step_avg:172.63ms
step:1329/1530 train_loss:3.2951 train_time:227862ms step_avg:172.75ms
step:1330/1530 train_loss:3.4362 train_time:228043ms step_avg:172.76ms
step:1331/1530 train_loss:3.3305 train_time:228220ms step_avg:172.76ms
step:1332/1530 train_loss:3.7445 train_time:228402ms step_avg:172.77ms
step:1333/1530 train_loss:3.4817 train_time:228583ms step_avg:172.78ms
step:1334/1530 train_loss:3.3721 train_time:228762ms step_avg:172.78ms
step:1335/1530 train_loss:3.2909 train_time:228941ms step_avg:172.79ms
step:1336/1530 train_loss:3.2973 train_time:229125ms step_avg:172.79ms
step:1337/1530 train_loss:3.5542 train_time:229305ms step_avg:172.80ms
step:1338/1530 train_loss:3.5273 train_time:229483ms step_avg:172.80ms
step:1339/1530 train_loss:3.3431 train_time:229662ms step_avg:172.81ms
step:1340/1530 train_loss:3.2868 train_time:229841ms step_avg:172.81ms
step:1341/1530 train_loss:3.5915 train_time:230019ms step_avg:172.82ms
step:1342/1530 train_loss:3.3632 train_time:230201ms step_avg:172.82ms
step:1343/1530 train_loss:3.3665 train_time:230378ms step_avg:172.83ms
step:1344/1530 train_loss:3.4215 train_time:230559ms step_avg:172.83ms
step:1345/1530 train_loss:3.3807 train_time:230742ms step_avg:172.84ms
step:1346/1530 train_loss:3.2999 train_time:230919ms step_avg:172.84ms
step:1347/1530 train_loss:3.2805 train_time:231096ms step_avg:172.85ms
step:1348/1530 train_loss:3.3517 train_time:231273ms step_avg:172.85ms
step:1349/1530 train_loss:3.2760 train_time:231449ms step_avg:172.85ms
step:1350/1530 train_loss:3.3938 train_time:231630ms step_avg:172.86ms
step:1351/1530 train_loss:3.2506 train_time:231806ms step_avg:172.86ms
step:1352/1530 train_loss:3.3085 train_time:231983ms step_avg:172.86ms
step:1353/1530 train_loss:3.4040 train_time:232163ms step_avg:172.87ms
step:1354/1530 train_loss:3.2618 train_time:232342ms step_avg:172.87ms
step:1355/1530 train_loss:3.1909 train_time:232518ms step_avg:172.88ms
step:1356/1530 train_loss:3.5097 train_time:232699ms step_avg:172.88ms
step:1357/1530 train_loss:3.4252 train_time:232879ms step_avg:172.89ms
step:1358/1530 train_loss:3.1865 train_time:233057ms step_avg:172.89ms
step:1359/1530 train_loss:3.4417 train_time:233238ms step_avg:172.90ms
step:1360/1530 train_loss:3.3481 train_time:233419ms step_avg:172.90ms
step:1361/1530 train_loss:3.1311 train_time:233603ms step_avg:172.91ms
step:1362/1530 train_loss:3.3959 train_time:233785ms step_avg:172.92ms
step:1363/1530 train_loss:3.2853 train_time:233972ms step_avg:172.93ms
step:1364/1530 train_loss:3.3021 train_time:234149ms step_avg:172.93ms
step:1365/1530 train_loss:3.3166 train_time:234328ms step_avg:172.94ms
step:1366/1530 train_loss:3.4220 train_time:234508ms step_avg:172.94ms
step:1367/1530 train_loss:3.3966 train_time:234686ms step_avg:172.94ms
step:1368/1530 train_loss:3.3507 train_time:234866ms step_avg:172.95ms
step:1369/1530 train_loss:3.2732 train_time:235052ms step_avg:172.96ms
step:1370/1530 train_loss:3.6104 train_time:235233ms step_avg:172.97ms
step:1371/1530 train_loss:3.3162 train_time:235412ms step_avg:172.97ms
step:1372/1530 train_loss:3.3725 train_time:235596ms step_avg:172.98ms
step:1373/1530 train_loss:3.3719 train_time:235774ms step_avg:172.98ms
step:1374/1530 train_loss:3.1525 train_time:235955ms step_avg:172.99ms
step:1375/1530 train_loss:3.5425 train_time:236136ms step_avg:172.99ms
step:1375/1530 val_loss:3.3138 train_time:236187ms step_avg:173.03ms
step:1376/1530 train_loss:3.3515 train_time:236315ms step_avg:173.00ms
step:1377/1530 train_loss:3.4863 train_time:236493ms step_avg:173.00ms
step:1378/1530 train_loss:3.4748 train_time:236671ms step_avg:173.01ms
step:1379/1530 train_loss:3.1248 train_time:236851ms step_avg:173.01ms
step:1380/1530 train_loss:3.3155 train_time:237031ms step_avg:173.02ms
step:1381/1530 train_loss:3.7075 train_time:237214ms step_avg:173.02ms
step:1382/1530 train_loss:3.2166 train_time:237393ms step_avg:173.03ms
step:1383/1530 train_loss:3.3905 train_time:237572ms step_avg:173.03ms
step:1384/1530 train_loss:3.4782 train_time:237757ms step_avg:173.04ms
step:1385/1530 train_loss:3.4077 train_time:237932ms step_avg:173.04ms
step:1386/1530 train_loss:3.3422 train_time:238111ms step_avg:173.05ms
step:1387/1530 train_loss:3.2003 train_time:238290ms step_avg:173.05ms
step:1388/1530 train_loss:3.3457 train_time:238468ms step_avg:173.05ms
step:1389/1530 train_loss:3.3173 train_time:238650ms step_avg:173.06ms
step:1390/1530 train_loss:3.5722 train_time:238829ms step_avg:173.06ms
step:1391/1530 train_loss:3.2885 train_time:239009ms step_avg:173.07ms
step:1392/1530 train_loss:3.2873 train_time:239188ms step_avg:173.07ms
step:1393/1530 train_loss:3.2415 train_time:239368ms step_avg:173.08ms
step:1394/1530 train_loss:3.4998 train_time:239545ms step_avg:173.08ms
step:1395/1530 train_loss:3.3947 train_time:239725ms step_avg:173.09ms
step:1396/1530 train_loss:3.4088 train_time:239902ms step_avg:173.09ms
step:1397/1530 train_loss:3.3154 train_time:240078ms step_avg:173.09ms
step:1398/1530 train_loss:3.2550 train_time:240254ms step_avg:173.09ms
step:1399/1530 train_loss:3.3150 train_time:240433ms step_avg:173.10ms
step:1400/1530 train_loss:3.3223 train_time:240617ms step_avg:173.11ms
step:1401/1530 train_loss:3.3544 train_time:240792ms step_avg:173.11ms
step:1402/1530 train_loss:3.3005 train_time:240971ms step_avg:173.11ms
step:1403/1530 train_loss:3.4963 train_time:241153ms step_avg:173.12ms
step:1404/1530 train_loss:3.2828 train_time:241330ms step_avg:173.12ms
step:1405/1530 train_loss:3.3179 train_time:241511ms step_avg:173.13ms
step:1406/1530 train_loss:3.3161 train_time:241691ms step_avg:173.13ms
step:1407/1530 train_loss:3.1756 train_time:241866ms step_avg:173.13ms
step:1408/1530 train_loss:3.3156 train_time:242046ms step_avg:173.14ms
step:1409/1530 train_loss:3.3013 train_time:242232ms step_avg:173.15ms
step:1410/1530 train_loss:3.2893 train_time:242408ms step_avg:173.15ms
step:1411/1530 train_loss:3.3642 train_time:242584ms step_avg:173.15ms
step:1412/1530 train_loss:3.3373 train_time:242761ms step_avg:173.15ms
step:1413/1530 train_loss:3.3651 train_time:242940ms step_avg:173.16ms
step:1414/1530 train_loss:3.3288 train_time:243120ms step_avg:173.16ms
step:1415/1530 train_loss:3.4102 train_time:243303ms step_avg:173.17ms
step:1416/1530 train_loss:3.2307 train_time:243491ms step_avg:173.18ms
step:1417/1530 train_loss:3.2832 train_time:243672ms step_avg:173.19ms
step:1418/1530 train_loss:3.3893 train_time:243853ms step_avg:173.19ms
step:1419/1530 train_loss:3.3485 train_time:244036ms step_avg:173.20ms
step:1420/1530 train_loss:3.3709 train_time:244216ms step_avg:173.20ms
step:1421/1530 train_loss:3.3751 train_time:244394ms step_avg:173.21ms
step:1422/1530 train_loss:3.3322 train_time:244571ms step_avg:173.21ms
step:1423/1530 train_loss:3.3182 train_time:244750ms step_avg:173.21ms
step:1424/1530 train_loss:3.3353 train_time:244936ms step_avg:173.22ms
step:1425/1530 train_loss:3.1922 train_time:245125ms step_avg:173.23ms
step:1426/1530 train_loss:3.3290 train_time:245304ms step_avg:173.24ms
step:1427/1530 train_loss:3.2861 train_time:245488ms step_avg:173.24ms
step:1428/1530 train_loss:3.3813 train_time:245666ms step_avg:173.25ms
step:1429/1530 train_loss:3.3535 train_time:245843ms step_avg:173.25ms
step:1430/1530 train_loss:3.2630 train_time:246026ms step_avg:173.26ms
step:1431/1530 train_loss:3.3222 train_time:246209ms step_avg:173.26ms
step:1432/1530 train_loss:3.3394 train_time:246390ms step_avg:173.27ms
step:1433/1530 train_loss:3.1335 train_time:246573ms step_avg:173.28ms
step:1434/1530 train_loss:3.2907 train_time:246756ms step_avg:173.28ms
step:1435/1530 train_loss:3.1169 train_time:246936ms step_avg:173.29ms
step:1436/1530 train_loss:3.2339 train_time:247117ms step_avg:173.29ms
step:1437/1530 train_loss:3.4087 train_time:247295ms step_avg:173.30ms
step:1438/1530 train_loss:3.3835 train_time:247470ms step_avg:173.30ms
step:1439/1530 train_loss:3.3142 train_time:247650ms step_avg:173.30ms
step:1440/1530 train_loss:3.1941 train_time:247826ms step_avg:173.30ms
step:1441/1530 train_loss:3.3422 train_time:248006ms step_avg:173.31ms
step:1442/1530 train_loss:3.3867 train_time:248189ms step_avg:173.32ms
step:1443/1530 train_loss:3.4942 train_time:248375ms step_avg:173.33ms
step:1444/1530 train_loss:3.4498 train_time:248552ms step_avg:173.33ms
step:1445/1530 train_loss:3.3371 train_time:248730ms step_avg:173.33ms
step:1446/1530 train_loss:3.1981 train_time:248910ms step_avg:173.34ms
step:1447/1530 train_loss:3.2984 train_time:249090ms step_avg:173.34ms
step:1448/1530 train_loss:3.2989 train_time:249267ms step_avg:173.34ms
step:1449/1530 train_loss:3.3997 train_time:249447ms step_avg:173.35ms
step:1450/1530 train_loss:3.3902 train_time:249629ms step_avg:173.35ms
step:1451/1530 train_loss:3.2091 train_time:249808ms step_avg:173.36ms
step:1452/1530 train_loss:3.3281 train_time:249988ms step_avg:173.36ms
step:1453/1530 train_loss:3.2589 train_time:250164ms step_avg:173.36ms
step:1454/1530 train_loss:3.2907 train_time:250341ms step_avg:173.37ms
step:1455/1530 train_loss:3.3267 train_time:250524ms step_avg:173.37ms
step:1456/1530 train_loss:3.2863 train_time:250703ms step_avg:173.38ms
step:1457/1530 train_loss:3.1572 train_time:250880ms step_avg:173.38ms
step:1458/1530 train_loss:3.4256 train_time:251059ms step_avg:173.38ms
step:1459/1530 train_loss:3.2695 train_time:251240ms step_avg:173.39ms
step:1460/1530 train_loss:3.3165 train_time:251419ms step_avg:173.39ms
step:1461/1530 train_loss:3.4300 train_time:251598ms step_avg:173.40ms
step:1462/1530 train_loss:3.2621 train_time:251775ms step_avg:173.40ms
step:1463/1530 train_loss:3.4710 train_time:251958ms step_avg:173.41ms
step:1464/1530 train_loss:3.3625 train_time:252135ms step_avg:173.41ms
step:1465/1530 train_loss:3.3609 train_time:252316ms step_avg:173.41ms
step:1466/1530 train_loss:3.2887 train_time:252492ms step_avg:173.42ms
step:1467/1530 train_loss:3.3967 train_time:252671ms step_avg:173.42ms
step:1468/1530 train_loss:3.2906 train_time:252847ms step_avg:173.42ms
step:1469/1530 train_loss:3.2772 train_time:253027ms step_avg:173.43ms
step:1470/1530 train_loss:3.3335 train_time:253209ms step_avg:173.43ms
step:1471/1530 train_loss:3.2623 train_time:253395ms step_avg:173.44ms
step:1472/1530 train_loss:3.2539 train_time:253580ms step_avg:173.45ms
step:1473/1530 train_loss:3.4436 train_time:253758ms step_avg:173.45ms
step:1474/1530 train_loss:3.3171 train_time:253941ms step_avg:173.46ms
step:1475/1530 train_loss:3.1522 train_time:254127ms step_avg:173.47ms
step:1476/1530 train_loss:3.2692 train_time:254305ms step_avg:173.47ms
step:1477/1530 train_loss:3.2415 train_time:254490ms step_avg:173.48ms
step:1478/1530 train_loss:3.3059 train_time:254674ms step_avg:173.48ms
step:1479/1530 train_loss:3.4001 train_time:254856ms step_avg:173.49ms
step:1480/1530 train_loss:3.2719 train_time:255033ms step_avg:173.49ms
step:1481/1530 train_loss:3.4538 train_time:255213ms step_avg:173.50ms
step:1482/1530 train_loss:3.3667 train_time:255402ms step_avg:173.51ms
step:1483/1530 train_loss:3.2781 train_time:255593ms step_avg:173.52ms
step:1484/1530 train_loss:3.2687 train_time:255779ms step_avg:173.53ms
step:1485/1530 train_loss:3.2844 train_time:255960ms step_avg:173.53ms
step:1486/1530 train_loss:3.2288 train_time:256147ms step_avg:173.54ms
step:1487/1530 train_loss:3.3456 train_time:256330ms step_avg:173.55ms
step:1488/1530 train_loss:3.2484 train_time:256514ms step_avg:173.55ms
step:1489/1530 train_loss:3.3171 train_time:256693ms step_avg:173.56ms
step:1490/1530 train_loss:3.2523 train_time:256872ms step_avg:173.56ms
step:1491/1530 train_loss:3.1594 train_time:257052ms step_avg:173.57ms
step:1492/1530 train_loss:3.2664 train_time:257232ms step_avg:173.57ms
step:1493/1530 train_loss:3.4356 train_time:257410ms step_avg:173.57ms
step:1494/1530 train_loss:3.3001 train_time:257588ms step_avg:173.58ms
step:1495/1530 train_loss:3.0337 train_time:257773ms step_avg:173.58ms
step:1496/1530 train_loss:3.3626 train_time:257954ms step_avg:173.59ms
step:1497/1530 train_loss:3.3169 train_time:258139ms step_avg:173.60ms
step:1498/1530 train_loss:3.3471 train_time:258323ms step_avg:173.60ms
step:1499/1530 train_loss:3.3155 train_time:258510ms step_avg:173.61ms
step:1500/1530 train_loss:3.2966 train_time:258701ms step_avg:173.62ms
step:1500/1530 val_loss:3.2818 train_time:258755ms step_avg:173.66ms
step:1501/1530 train_loss:3.0888 train_time:258891ms step_avg:173.64ms
step:1502/1530 train_loss:3.3644 train_time:259081ms step_avg:173.65ms
step:1503/1530 train_loss:3.2469 train_time:259259ms step_avg:173.65ms
step:1504/1530 train_loss:3.2515 train_time:259439ms step_avg:173.65ms
step:1505/1530 train_loss:3.2153 train_time:259619ms step_avg:173.66ms
step:1506/1530 train_loss:3.2813 train_time:259802ms step_avg:173.66ms
step:1507/1530 train_loss:3.1804 train_time:259997ms step_avg:173.68ms
step:1508/1530 train_loss:3.4845 train_time:260180ms step_avg:173.68ms
step:1509/1530 train_loss:3.2857 train_time:260357ms step_avg:173.69ms
step:1510/1530 train_loss:3.2727 train_time:260538ms step_avg:173.69ms
step:1511/1530 train_loss:3.4163 train_time:260850ms step_avg:173.78ms
step:1512/1530 train_loss:3.4215 train_time:261038ms step_avg:173.79ms
step:1513/1530 train_loss:3.2737 train_time:261222ms step_avg:173.80ms
step:1514/1530 train_loss:3.0855 train_time:261404ms step_avg:173.81ms
step:1515/1530 train_loss:3.2465 train_time:261584ms step_avg:173.81ms
step:1516/1530 train_loss:3.2575 train_time:261770ms step_avg:173.82ms
step:1517/1530 train_loss:3.3023 train_time:261951ms step_avg:173.82ms
step:1518/1530 train_loss:3.2087 train_time:262135ms step_avg:173.83ms
step:1519/1530 train_loss:3.5000 train_time:262464ms step_avg:173.93ms
step:1520/1530 train_loss:3.1313 train_time:262649ms step_avg:173.94ms
step:1521/1530 train_loss:3.2050 train_time:262826ms step_avg:173.94ms
step:1522/1530 train_loss:3.3587 train_time:263010ms step_avg:173.95ms
step:1523/1530 train_loss:3.2346 train_time:263189ms step_avg:173.95ms
step:1524/1530 train_loss:3.3503 train_time:263370ms step_avg:173.96ms
step:1525/1530 train_loss:3.3388 train_time:263558ms step_avg:173.97ms
step:1526/1530 train_loss:3.2758 train_time:263748ms step_avg:173.98ms
step:1527/1530 train_loss:3.2911 train_time:263928ms step_avg:173.98ms
step:1528/1530 train_loss:3.4124 train_time:264107ms step_avg:173.98ms
step:1529/1530 train_loss:3.4100 train_time:264284ms step_avg:173.99ms
step:1530/1530 train_loss:3.2414 train_time:264461ms step_avg:173.99ms
step:1530/1530 val_loss:3.2793 train_time:264515ms step_avg:174.02ms