Qwen3-1.7B-Base-nq-baseline-scratch

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

<details><summary>See axolotl config</summary>

axolotl version: 0.10.0

base_model: Qwen/Qwen3-1.7B-Base
datasets:
  - path: Abner0803/nq_text-with_pseudo_query-100k-gr
    data_files: data/icl_test.jsonl
    type: chat_template
    chat_template: tokenizer_default_fallback_chatml
    field_messages: conversations
    message_property_mappings:
      role: role
      content: content
    roles:
      assistant:
        - assistant
        - gpt
        - model
      user:
        - user
        - human
      system:
        - system
    roles_to_train: ["assistant"]
    train_on_eos: "turn"

dataset_processes: 6
streaming: false
shuffle_merged_datasets: true
output_dir: /home/theblackcat/ICLGR/checkpoints/Qwen3-1.7B-Base-nq-baseline-scratch
sequence_len: 2048
sample_packing: false
flash_attention: false
xformers_attention: false
flex_attention: false
sdp_attention: true
overrides_of_model_config:
  _attn_implementation: "sdpa"

gradient_accumulation_steps: 128
micro_batch_size: 1
num_epochs: 10

optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 0.0001
warmup_ratio: 0.1
weight_decay: 0.0
bf16: true
tf32: false
gradient_checkpointing: true

logging_steps: 50
save_strategy: steps
save_steps: 1000
save_total_limit: 3

special_tokens:
  eos_token: "<|im_end|>"

val_set_size: 0.0
wandb_project: ICLGR-NQ
wandb_entity: abnerden0803-national-taiwan-university
wandb_watch:
wandb_name: qwen3-1.7b-base-nq-baseline-scratch
wandb_log_model:

</details><br>

home/theblackcat/ICLGR/checkpoints/Qwen3-1.7B-Base-nq-baseline-scratch

This model is a fine-tuned version of Qwen/Qwen3-1.7B-Base on the Abner0803/nq_text-with_pseudo_query-100k-gr dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 128
• total_train_batch_size: 128
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 412
• training_steps: 4123

Training results

Framework versions

• Transformers 4.52.3
• Pytorch 2.9.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.4