laurent-maille/testLMA

See axolotl config

axolotl version: 0.8.0.dev0

base_model: Qwen/Qwen3-4B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
adapter: lora
lora_r: 32
lora_alpha: 32
lora_dropout: 0.05
lora_target_modules:
- q_proj
- k_proj
- v_proj
- o_proj
- gate_proj
- up_proj
- down_proj
datasets:
- path: laurent-maille/plc-draft2
  type: chat_template
  field_messages: messages
  conversation: chat
dataset_prepared_path: ./prepared/plc_sharegpt
output_dir: ./outputs/ax_qwen3_4b_8bit
sequence_len: 4096
sample_packing: true
gradient_checkpointing: true
use_flash_attention: true
bf16: true
tf32: true
seed: 42
optimizer: adamw_bnb_8bit
learning_rate: 0.0001
weight_decay: 0.0
lr_scheduler: cosine
warmup_ratio: 0.03
micro_batch_size: 1
gradient_accumulation_steps: 8
num_epochs: 1
logging_steps: 10
save_strategy: steps
save_steps: 500
eval_steps: 0
evaluation_strategy: 'no'
cutoff_len: 4096
flash_attn_impl: fa2
load_in_8bit: true
fp16: false
packing: true
pad_to_sequence_len: true
group_by_length: true
max_grad_norm: 0.3
save_safetensors: true
save_total_limit: 2

outputs/ax_qwen3_4b_8bit

This model is a fine-tuned version of Qwen/Qwen3-4B on the laurent-maille/plc-draft2 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: 8
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 1.0

Training results

Framework versions

• PEFT 0.14.0
• Transformers 4.55.4
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0