step 50

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "gpt2",
+  "_name_or_path": "./",
   "activation_function": "gelu_new",
   "architectures": [
     "GPT2LMHeadModel"

gptesla_training.py

@@ -0,0 +1,288 @@
+import os
+
+import datasets, transformers
+from transformers import AutoConfig, AutoModelForCausalLM, AutoTokenizer, set_seed
+from transformers.optimization import get_scheduler
+from datasets import load_dataset, DownloadConfig
+
+import torch
+from torch.utils.data import IterableDataset
+from torch.utils.data.dataloader import DataLoader
+from torch.utils.tensorboard import SummaryWriter
+from torch.optim import AdamW
+
+import logging
+import wandb
+from huggingface_hub import Repository, create_branch
+from accelerate import Accelerator
+from argparse import Namespace
+
+
+# Set the API token as an environment variable
+os.environ["TOKENIZERS_PARALLELISM"] = "false"
+
+
+def save_checkpoint_state():
+
+    dir_name = "./torch_checkpoint"
+    os.makedirs(dir_name, exist_ok=True)
+
+    checkpoint = {
+        "lr_scheduler": lr_scheduler.state_dict(),
+        "completed_steps": completed_steps,
+        "run_name": run_name,
+        "optimizer": optimizer.state_dict(),
+        "run_id": wandb_id
+    }
+    torch.save(checkpoint, f"torch_checkpoint/latest_checkpoint.pth")
+
+
+class ConstantLengthDataset(IterableDataset):
+
+    def __init__(
+        self,
+        tokenizer,
+        dataset,
+        seq_length=1024,
+        num_of_sequences=1024,
+        chars_per_token=3.6,
+    ):
+        self.tokenizer = tokenizer
+        self.concat_token_id = tokenizer.eos_token_id
+        self.dataset = dataset
+        self.seq_length = seq_length
+        self.input_characters = seq_length * chars_per_token * num_of_sequences
+
+    def __iter__(self):
+        iterator = iter(self.dataset)
+        more_examples = True
+        while more_examples:
+            buffer, buffer_len = [], 0
+            while True:
+                if buffer_len >= self.input_characters:
+                    m = f"Buffer full: {buffer_len}>={self.input_characters:.0f}"
+                    # print(m)
+                    break
+                try:
+                    m = f"Fill buffer: {buffer_len}<{self.input_characters:.0f}"
+                    # print(m)
+                    buffer.append(next(iterator)["content"])
+                    buffer_len += len(buffer[-1])
+                except StopIteration:
+                    # iterator = iter(self.dataset)
+                    more_examples = False
+                    break
+
+            all_token_ids = []
+            tokenized_inputs = self.tokenizer(buffer, truncation=False)
+            for tokenized_input in tokenized_inputs["input_ids"]:
+                all_token_ids.extend(tokenized_input + [self.concat_token_id])
+
+            for i in range(0, len(all_token_ids), self.seq_length):
+                input_ids = all_token_ids[i : i + self.seq_length]
+                if len(input_ids) == self.seq_length:
+                    yield torch.tensor(input_ids)
+
+
+def setup_logging(project_name):
+    logger = logging.getLogger(__name__)
+
+    dir_name = "./log"
+    if not os.path.exists(dir_name):
+        os.makedirs(dir_name)
+        print(f"Directory '{dir_name}' was created.")
+    else:
+        print(f"Directory '{dir_name}' already exists.")
+
+    # setting up log directory
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        datefmt="%m/%d/%Y %H:%M:%S",
+        level=logging.INFO,
+        handlers=[
+            logging.FileHandler(f"log/debug_{accelerator.process_index}.log"),
+            logging.StreamHandler(),
+        ],
+    )
+    if accelerator.is_main_process:  # We only want to set up logging once
+        wandb.init(project=project_name, config=args, dir="./../")
+        run_name = wandb.run.name
+        wandb_id = wandb.run.id
+        tb_writer = SummaryWriter()
+        tb_writer.add_hparams(vars(args), {"0": 0})
+        logger.setLevel(logging.INFO)
+        datasets.utils.logging.set_verbosity_debug()
+        transformers.utils.logging.set_verbosity_info()
+    else:
+        tb_writer = None
+        run_name = ""
+        wandb_id = ""
+        logger.setLevel(logging.ERROR)
+        datasets.utils.logging.set_verbosity_error()
+        transformers.utils.logging.set_verbosity_error()
+    return logger, tb_writer, run_name, wandb_id
+
+
+def create_dataloaders(dataset_name):
+    train_data = load_dataset(dataset_name + "-train", split="train", streaming=True)
+    train_data = train_data.shuffle(buffer_size=args.shuffle_buffer, seed=args.seed)
+    valid_data = load_dataset(dataset_name + "-valid", split="validation", streaming=True)
+
+    train_dataset = ConstantLengthDataset(tokenizer, train_data, seq_length=args.seq_length)
+    valid_dataset = ConstantLengthDataset(tokenizer, valid_data, seq_length=args.seq_length)
+
+    train_dataloader = DataLoader(train_dataset, batch_size=args.train_batch_size, num_workers=96)
+    eval_dataloader = DataLoader(valid_dataset, batch_size=args.valid_batch_size, num_workers=1)
+    return train_dataloader, eval_dataloader
+
+
+def log_metrics(step, metrics):
+    logger.info(f"Step {step}: {metrics}")
+    if accelerator.is_main_process:
+        wandb.log(metrics)
+        [tb_writer.add_scalar(k, v, step) for k, v in metrics.items()]
+
+
+def get_grouped_params(model, no_decay=["bias", "LayerNorm.weight"]):
+    params_with_wd, params_without_wd = [], []
+    for n, p in model.named_parameters():
+        if any(nd in n for nd in no_decay):
+            params_without_wd.append(p)
+        else:
+            params_with_wd.append(p)
+    return [
+        {"params": params_with_wd, "weight_decay": args.weight_decay},
+        {"params": params_without_wd, "weight_decay": 0.0},
+    ]
+
+
+def evaluate():
+    model.eval()
+    losses = []
+    for step, batch in enumerate(eval_dataloader):
+        with torch.no_grad():
+            outputs = model(batch, labels=batch)
+        loss = outputs.loss.repeat(args.valid_batch_size)
+        losses.append(accelerator.gather(loss))
+        if args.max_eval_steps > 0 and step >= args.max_eval_steps:
+            break
+    loss = torch.mean(torch.cat(losses))
+
+    try:
+        perplexity = torch.exp(loss)
+    except OverflowError:
+        perplexity = torch.tensor(float("inf"))
+
+    return loss.item(), perplexity.item()
+
+
+# Accelerator
+accelerator = Accelerator(dispatch_batches=True)
+acc_state = {str(k): str(v) for k, v in accelerator.state.__dict__.items()}
+
+project_name = "shng2025/gptesla-small"
+dataset_name = "shng2025/gptesla"
+
+# GPTesla - 111M param setup in comment. Modification to make lighter training requirement needed
+config = {
+    "train_batch_size": 12,  # 12
+    "valid_batch_size": 12,  # 12
+    "weight_decay": 0.1,
+    "shuffle_buffer": 1000,
+    "learning_rate": 5e-4,  # 5e-4
+    "lr_scheduler_type": "cosine",
+    "num_warmup_steps": 700,  # 2000
+    "gradient_accumulation_steps": 1,  # 1
+    "max_train_steps": 50000,  # 150000
+    "max_eval_steps": 10,
+    "seq_length": 1024,
+    "seed": 1,
+    "save_checkpoint_steps": 50,
+}  # 15000
+
+args = Namespace(**config, **acc_state)
+samples_per_step = accelerator.state.num_processes * args.train_batch_size
+set_seed(args.seed)
+
+# Logging
+logger, tb_writer, run_name, wandb_id = setup_logging(project_name.split("/")[1])
+logger.info(accelerator.state)
+
+# Load model and tokenizer
+if accelerator.is_main_process:
+    new_branch_name = run_name
+    create_branch(project_name, repo_type="model", branch=new_branch_name)
+    hf_repo = Repository("./", clone_from=project_name, revision=run_name)
+
+model = AutoModelForCausalLM.from_pretrained("./")  # , gradient_checkpointing=True)
+tokenizer = AutoTokenizer.from_pretrained("./")
+
+# Load dataset and dataloader
+train_dataloader, eval_dataloader = create_dataloaders(dataset_name)
+
+# Prepare the optimizer and learning rate scheduler
+optimizer = AdamW(get_grouped_params(model), lr=args.learning_rate)
+lr_scheduler = get_scheduler(
+    name=args.lr_scheduler_type,
+    optimizer=optimizer,
+    num_warmup_steps=args.num_warmup_steps,
+    num_training_steps=args.max_train_steps,
+)
+
+
+def get_lr():
+    return optimizer.param_groups[0]["lr"]
+
+
+# Prepare everything with our `accelerator` (order of args is not important)
+model, optimizer, train_dataloader, eval_dataloader = accelerator.prepare(
+    model, optimizer, train_dataloader, eval_dataloader
+)
+
+print(wandb_id)
+
+# Train model
+model.train()
+completed_steps = 0
+for step, batch in enumerate(train_dataloader, start=1):
+    loss = model(batch, labels=batch).loss
+    log_metrics(
+        step,
+        {
+            "lr": get_lr(),
+            "samples": step * samples_per_step,
+            "steps": completed_steps,
+            "loss/train": loss.item(),
+        },
+    )
+    loss = loss / args.gradient_accumulation_steps
+    accelerator.backward(loss)
+    if step % args.gradient_accumulation_steps == 0:
+        optimizer.step()
+        lr_scheduler.step()
+        optimizer.zero_grad()
+        completed_steps += 1
+    if step % args.save_checkpoint_steps == 0:
+        logger.info("Evaluating and saving model checkpoint")
+        eval_loss, perplexity = evaluate()
+        log_metrics(step, {"loss/eval": eval_loss, "perplexity": perplexity})
+        accelerator.wait_for_everyone()
+        unwrapped_model = accelerator.unwrap_model(model)
+        if accelerator.is_main_process:
+            save_checkpoint_state()
+            unwrapped_model.save_pretrained("./")
+            hf_repo.push_to_hub(commit_message=f"step {step}")
+        model.train()
+    if completed_steps >= args.max_train_steps:
+        break
+
+
+# Evaluate and save the last checkpoint
+logger.info("Evaluating and saving model after training")
+eval_loss, perplexity = evaluate()
+log_metrics(step, {"loss/eval": eval_loss, "perplexity": perplexity})
+accelerator.wait_for_everyone()
+unwrapped_model = accelerator.unwrap_model(model)
+if accelerator.is_main_process:
+    unwrapped_model.save_pretrained("./")
+    hf_repo.push_to_hub(commit_message="final model")

log/debug_0.log

@@ -0,0 +1,243 @@
+07/25/2024 10:33:00 - INFO - __main__ - Distributed environment: MULTI_GPU  Backend: nccl
+Num processes: 4
+Process index: 0
+Local process index: 0
+Device: cuda:0
+
+Mixed precision type: fp16
+
+07/25/2024 10:33:01 - WARNING - huggingface_hub.repository - /dli/gptesla-small/./ is already a clone of https://huggingface.co/shng2025/gptesla-small. Make sure you pull the latest changes with `repo.git_pull()`.
+07/25/2024 10:33:01 - WARNING - huggingface_hub.repository - Revision `fiery-sea-141` does not exist. Created and checked out branch `fiery-sea-141`.
+07/25/2024 10:33:01 - WARNING - huggingface_hub.repository - 
+07/25/2024 10:33:02 - DEBUG - datasets.utils._dataset_viewer - Dataset info for shng2025/gptesla-train is not completely ready yet.
+07/25/2024 10:33:03 - INFO - datasets.builder - No config specified, defaulting to the single config: gptesla-train/default
+07/25/2024 10:33:03 - INFO - datasets.info - Loading Dataset Infos from /usr/local/lib/python3.10/dist-packages/datasets/packaged_modules/json
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#0, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#2, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#1, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#3, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#4, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#6, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#7, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#8, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#5, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#9, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#11, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#14, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#15, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#13, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#17, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#18, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#19, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#20, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#16, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#12, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#10, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#21, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#24, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#22, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#23, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#25, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#26, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#27, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#28, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#29, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#30, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#31, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#35, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#33, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#37, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#32, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:07 - DEBUG - datasets.iterable_dataset - dataloader worker#34, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#36, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#38, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#41, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#39, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#42, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#40, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#43, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#46, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#45, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#47, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#51, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#49, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#54, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#44, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#50, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#52, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#53, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#55, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#48, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#56, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#58, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#57, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#60, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#59, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#61, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#62, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#63, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#65, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#64, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#66, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#68, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#67, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#69, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#70, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#71, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#72, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#73, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#74, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#75, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#76, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#78, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#77, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#79, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#80, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#81, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#82, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#84, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#83, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#85, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#86, ': Starting to iterate over 2/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#87, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#88, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#89, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#90, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#91, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#93, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#92, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#94, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.iterable_dataset - dataloader worker#95, ': Starting to iterate over 1/183 shards.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486397 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492861 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489635 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486023 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492277 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10668116 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489599 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486023 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492861 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10522596 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491327 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497111 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497218 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10500930 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10512203 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485912 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497111 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10525688 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485912 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10562022 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10536479 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495973 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10598254 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489575 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10499106 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10499607 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491889 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487790 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488608 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10511604 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491272 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497062 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485842 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10552417 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10493913 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488150 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10552417 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485847 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509286 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10949076 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10863935 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486616 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491547 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 11286262 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488098 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509286 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 11286262 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486616 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10511500 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10525926 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488385 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488651 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:08 - DEBUG - datasets.packaged_modules.json.json - Batch of 10501535 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10515063 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488651 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10553677 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10511515 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487097 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10553677 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486276 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486172 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497335 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486801 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509262 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487725 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10610581 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10686322 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10621496 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10676628 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485918 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487482 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10500290 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10498167 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492554 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495520 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10640425 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495520 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 11115863 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10530453 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
+07/25/2024 10:33:09 - DEBUG - datasets.packaged_modules.json.json - Batch of 10751338 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
+07/25/2024 10:33:22 - INFO - __main__ - Step 1: {'lr': 0.0, 'samples': 48, 'steps': 0, 'loss/train': 10.554669380187988}
+07/25/2024 10:33:23 - INFO - __main__ - Step 2: {'lr': 7.142857142857143e-07, 'samples': 96, 'steps': 1, 'loss/train': 10.494059562683105}
+07/25/2024 10:33:24 - INFO - __main__ - Step 3: {'lr': 1.4285714285714286e-06, 'samples': 144, 'steps': 2, 'loss/train': 10.507988929748535}
+07/25/2024 10:33:24 - INFO - __main__ - Step 4: {'lr': 2.142857142857143e-06, 'samples': 192, 'steps': 3, 'loss/train': 10.415447235107422}
+07/25/2024 10:33:24 - INFO - __main__ - Step 5: {'lr': 2.8571428571428573e-06, 'samples': 240, 'steps': 4, 'loss/train': 10.345850944519043}
+07/25/2024 10:33:24 - INFO - __main__ - Step 6: {'lr': 3.5714285714285714e-06, 'samples': 288, 'steps': 5, 'loss/train': 10.195524215698242}
+07/25/2024 10:33:25 - INFO - __main__ - Step 7: {'lr': 4.285714285714286e-06, 'samples': 336, 'steps': 6, 'loss/train': 10.09341812133789}
+07/25/2024 10:33:25 - INFO - __main__ - Step 8: {'lr': 5e-06, 'samples': 384, 'steps': 7, 'loss/train': 9.965239524841309}
+07/25/2024 10:33:25 - INFO - __main__ - Step 9: {'lr': 5.7142857142857145e-06, 'samples': 432, 'steps': 8, 'loss/train': 9.698853492736816}
+07/25/2024 10:33:25 - INFO - __main__ - Step 10: {'lr': 6.428571428571429e-06, 'samples': 480, 'steps': 9, 'loss/train': 9.80683708190918}
+07/25/2024 10:33:26 - INFO - __main__ - Step 11: {'lr': 7.142857142857143e-06, 'samples': 528, 'steps': 10, 'loss/train': 9.633079528808594}
+07/25/2024 10:33:26 - INFO - __main__ - Step 12: {'lr': 7.857142857142858e-06, 'samples': 576, 'steps': 11, 'loss/train': 9.700591087341309}
+07/25/2024 10:33:26 - INFO - __main__ - Step 13: {'lr': 8.571428571428573e-06, 'samples': 624, 'steps': 12, 'loss/train': 9.603139877319336}
+07/25/2024 10:33:27 - INFO - __main__ - Step 14: {'lr': 9.285714285714286e-06, 'samples': 672, 'steps': 13, 'loss/train': 9.30308723449707}
+07/25/2024 10:33:27 - INFO - __main__ - Step 15: {'lr': 1e-05, 'samples': 720, 'steps': 14, 'loss/train': 9.333526611328125}
+07/25/2024 10:33:27 - INFO - __main__ - Step 16: {'lr': 1.0714285714285714e-05, 'samples': 768, 'steps': 15, 'loss/train': 8.336181640625}
+07/25/2024 10:33:27 - INFO - __main__ - Step 17: {'lr': 1.1428571428571429e-05, 'samples': 816, 'steps': 16, 'loss/train': 9.075631141662598}
+07/25/2024 10:33:28 - INFO - __main__ - Step 18: {'lr': 1.2142857142857142e-05, 'samples': 864, 'steps': 17, 'loss/train': 9.18478012084961}
+07/25/2024 10:33:28 - INFO - __main__ - Step 19: {'lr': 1.2857142857142857e-05, 'samples': 912, 'steps': 18, 'loss/train': 8.96328353881836}
+07/25/2024 10:33:28 - INFO - __main__ - Step 20: {'lr': 1.3571428571428572e-05, 'samples': 960, 'steps': 19, 'loss/train': 9.45018196105957}
+07/25/2024 10:33:29 - INFO - __main__ - Step 21: {'lr': 1.4285714285714285e-05, 'samples': 1008, 'steps': 20, 'loss/train': 8.517333984375}
+07/25/2024 10:33:29 - INFO - __main__ - Step 22: {'lr': 1.5e-05, 'samples': 1056, 'steps': 21, 'loss/train': 9.207684516906738}
+07/25/2024 10:33:29 - INFO - __main__ - Step 23: {'lr': 1.5714285714285715e-05, 'samples': 1104, 'steps': 22, 'loss/train': 8.681092262268066}
+07/25/2024 10:33:29 - INFO - __main__ - Step 24: {'lr': 1.642857142857143e-05, 'samples': 1152, 'steps': 23, 'loss/train': 8.316036224365234}
+07/25/2024 10:33:30 - INFO - __main__ - Step 25: {'lr': 1.7142857142857145e-05, 'samples': 1200, 'steps': 24, 'loss/train': 8.944169044494629}
+07/25/2024 10:33:30 - INFO - __main__ - Step 26: {'lr': 1.7857142857142855e-05, 'samples': 1248, 'steps': 25, 'loss/train': 8.878201484680176}
+07/25/2024 10:33:30 - INFO - __main__ - Step 27: {'lr': 1.8571428571428572e-05, 'samples': 1296, 'steps': 26, 'loss/train': 9.158102989196777}
+07/25/2024 10:33:30 - INFO - __main__ - Step 28: {'lr': 1.9285714285714285e-05, 'samples': 1344, 'steps': 27, 'loss/train': 9.14354419708252}
+07/25/2024 10:33:31 - INFO - __main__ - Step 29: {'lr': 2e-05, 'samples': 1392, 'steps': 28, 'loss/train': 8.860624313354492}
+07/25/2024 10:33:31 - INFO - __main__ - Step 30: {'lr': 2.0714285714285715e-05, 'samples': 1440, 'steps': 29, 'loss/train': 8.876450538635254}
+07/25/2024 10:33:31 - INFO - __main__ - Step 31: {'lr': 2.1428571428571428e-05, 'samples': 1488, 'steps': 30, 'loss/train': 8.425738334655762}
+07/25/2024 10:33:32 - INFO - __main__ - Step 32: {'lr': 2.214285714285714e-05, 'samples': 1536, 'steps': 31, 'loss/train': 8.942279815673828}
+07/25/2024 10:33:32 - INFO - __main__ - Step 33: {'lr': 2.2857142857142858e-05, 'samples': 1584, 'steps': 32, 'loss/train': 8.757084846496582}
+07/25/2024 10:33:32 - INFO - __main__ - Step 34: {'lr': 2.3571428571428575e-05, 'samples': 1632, 'steps': 33, 'loss/train': 8.699286460876465}
+07/25/2024 10:33:32 - INFO - __main__ - Step 35: {'lr': 2.4285714285714285e-05, 'samples': 1680, 'steps': 34, 'loss/train': 8.857367515563965}
+07/25/2024 10:33:33 - INFO - __main__ - Step 36: {'lr': 2.5e-05, 'samples': 1728, 'steps': 35, 'loss/train': 8.830195426940918}
+07/25/2024 10:33:33 - INFO - __main__ - Step 37: {'lr': 2.5714285714285714e-05, 'samples': 1776, 'steps': 36, 'loss/train': 8.944982528686523}
+07/25/2024 10:33:33 - INFO - __main__ - Step 38: {'lr': 2.642857142857143e-05, 'samples': 1824, 'steps': 37, 'loss/train': 8.670278549194336}
+07/25/2024 10:33:34 - INFO - __main__ - Step 39: {'lr': 2.7142857142857144e-05, 'samples': 1872, 'steps': 38, 'loss/train': 8.710525512695312}
+07/25/2024 10:33:34 - INFO - __main__ - Step 40: {'lr': 2.7857142857142858e-05, 'samples': 1920, 'steps': 39, 'loss/train': 7.902089595794678}
+07/25/2024 10:33:34 - INFO - __main__ - Step 41: {'lr': 2.857142857142857e-05, 'samples': 1968, 'steps': 40, 'loss/train': 8.400484085083008}
+07/25/2024 10:33:34 - INFO - __main__ - Step 42: {'lr': 2.9285714285714288e-05, 'samples': 2016, 'steps': 41, 'loss/train': 8.789310455322266}
+07/25/2024 10:33:35 - INFO - __main__ - Step 43: {'lr': 3e-05, 'samples': 2064, 'steps': 42, 'loss/train': 8.754344940185547}
+07/25/2024 10:33:35 - INFO - __main__ - Step 44: {'lr': 3.071428571428572e-05, 'samples': 2112, 'steps': 43, 'loss/train': 8.84192943572998}
+07/25/2024 10:33:35 - INFO - __main__ - Step 45: {'lr': 3.142857142857143e-05, 'samples': 2160, 'steps': 44, 'loss/train': 8.784793853759766}
+07/25/2024 10:33:35 - INFO - __main__ - Step 46: {'lr': 3.214285714285714e-05, 'samples': 2208, 'steps': 45, 'loss/train': 8.67403793334961}
+07/25/2024 10:33:36 - INFO - __main__ - Step 47: {'lr': 3.285714285714286e-05, 'samples': 2256, 'steps': 46, 'loss/train': 8.51427173614502}
+07/25/2024 10:33:36 - INFO - __main__ - Step 48: {'lr': 3.357142857142857e-05, 'samples': 2304, 'steps': 47, 'loss/train': 8.48193073272705}
+07/25/2024 10:33:36 - INFO - __main__ - Step 49: {'lr': 3.428571428571429e-05, 'samples': 2352, 'steps': 48, 'loss/train': 8.518038749694824}
+07/25/2024 10:33:37 - INFO - __main__ - Step 50: {'lr': 3.5000000000000004e-05, 'samples': 2400, 'steps': 49, 'loss/train': 8.63569450378418}
+07/25/2024 10:33:37 - INFO - __main__ - Evaluating and saving model checkpoint
+07/25/2024 10:33:37 - DEBUG - datasets.iterable_dataset - dataloader worker#0, ': Starting to iterate over 1/1 shards.
+07/25/2024 10:33:40 - INFO - __main__ - Step 50: {'loss/eval': 8.551246643066406, 'perplexity': 5173.19970703125}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e5bfbaf0ff37c9cbbbf37dfdf848aaee9b060a6a19f3c79d815806112c6c2f1
+oid sha256:7eb13d61f8d3f9cb945838abd62b274c030a665710e11a81e40523fd167676e8
 size 444048000

runs/Jul25_10-33-00_lab/1721903580.988387/events.out.tfevents.1721903580.lab.282621.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f95c37b42b28ae909301c5a3aaaa887964e7176e5ef00a6bfa378c26bd78842d
+size 1702

runs/Jul25_10-33-00_lab/events.out.tfevents.1721903580.lab.282621.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e6f77d111f92f23cc2b2a9b552566f4a3060821276e2d10780160b4addd9b34
+size 8983

torch_checkpoint/latest_checkpoint.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e95f68e321b4986d079008d4714de63f0608a42bd2ebabc6d695b6ac521c3001
+size 888193914