qwen80b-style-lora

<!-- 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: 0.13.0.dev0

adapter: lora
base_model: Qwen/Qwen2.5-72B-Instruct
load_in_4bit: true
bnb_4bit_compute_dtype: bfloat16
bnb_4bit_use_double_quant: true
bnb_4bit_quant_type: nf4

datasets:
  - path: ./patched_dataset/data.jsonl
    type: alpaca

val_set_size: 0.05
output_dir: ./outputs/qwen80b_qlora_run

micro_batch_size: 1
gradient_accumulation_steps: 8
num_epochs: 3
learning_rate: 2e-4

lora_alpha: 16
lora_r: 8
lora_dropout: 0.05
lora_target_modules:
  - q_proj
  - v_proj
  - k_proj
  - o_proj
  - gate_proj
  - down_proj
  - up_proj

sequence_len: 1024
train_on_inputs: false
optimizer: paged_adamw_8bit

bf16: true
fp16: false
tf32: true
gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false

warmup_ratio: 0.03
weight_decay: 0.01
logging_steps: 10
saves_per_epoch: 1
evals_per_epoch: 1
save_total_limit: 2

device_map: "auto"
low_cpu_mem_usage: true
torch_dtype: bfloat16

</details><br>

outputs/qwen80b_qlora_run

This model is a fine-tuned version of Qwen/Qwen2.5-72B-Instruct on the ./patched_dataset/data.jsonl dataset. It achieves the following results on the evaluation set:

• Loss: 1.8941
• Memory/max Active (gib): 43.77
• Memory/max Allocated (gib): 43.77
• Memory/device Reserved (gib): 45.94

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_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: 2
• training_steps: 90

Training results

Framework versions

• PEFT 0.17.1
• Transformers 4.57.0
• Pytorch 2.7.1+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1