Undi95/QwQ-RP-LoRA

See axolotl config

axolotl version: 0.8.0.dev0

base_model: ./Qwen_QwQ-32B/
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

trust_remote_code: true

load_in_8bit: true
load_in_4bit: false
strict: false

chat_template: tokenizer_default

datasets:
  - path: Undi95/QwQ-dataset
    type: chat_template
    chat_template: tokenizer_default
    field_messages: conversations
    message_field_role: from
    message_field_content: value
    roles:
      user: ["human", "user"]
      assistant: ["gpt", "assistant"]
      system: ["system"]
      tool: ["tool"]
dataset_prepared_path: last_run_prepared
val_set_size: 0.05
output_dir: ./out

sequence_len: 4096
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 256
lora_alpha: 64
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project: qwq-rp
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 2
micro_batch_size: 2
num_epochs: 2
optimizer: paged_adamw_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: true

gradient_checkpointing: unsloth
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 20
saves_per_epoch: 2
debug:
deepspeed:
weight_decay: 0.1

out

This model was trained from scratch on the Undi95/QwQ-dataset dataset. It achieves the following results on the evaluation set:

• Loss: 1.0077

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: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• total_eval_batch_size: 8
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT 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: 20
• num_epochs: 2.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.7216	1.0	649	1.0138
0.6349	1.9977	1296	1.0077

Framework versions

• PEFT 0.14.0
• Transformers 4.49.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0