qrk-labs/QRK-Mini-ALlM-Qwen3-1.7B-PEFT

See axolotl config

axolotl version: 0.9.1.post1

adapter: lora
base_model: Qwen/Qwen3-1.7B
bf16: auto
dataset_processes: 32
# Dataset configuration
datasets:
  - path: qrk-labs/QRK-Islam-Basic-Weak # <--- Your dataset on Hugging Face Hub
    type: chat_template # <--- Specify the chat template type
    drop_system_message: true
    field_messages: messages
    message_property_mappings: # <--- Included the message property mappings
      content: content
      role: role
    trust_remote_code: false # <--- Keep this false for security unless absolutely necessary
    conversation_template: chat_template # <--- Use the chatml template for Qwen models
gradient_accumulation_steps: 1
gradient_checkpointing: false
learning_rate: 0.0002
lisa_layers_attribute: model.layers
load_best_model_at_end: false
load_in_4bit: false
load_in_8bit: true
lora_alpha: 16
lora_dropout: 0.05
lora_r: 8
lora_target_modules:
- q_proj
- v_proj
- k_proj
- o_proj
- gate_proj
- down_proj
- up_proj
loraplus_lr_embedding: 1.0e-06
lr_scheduler: cosine
max_prompt_len: 512
mean_resizing_embeddings: false
micro_batch_size: 16
num_epochs: 1.0
optimizer: adamw_bnb_8bit
output_dir: ./outputs/mymodel
pretrain_multipack_attn: true
pretrain_multipack_buffer_size: 10000
qlora_sharded_model_loading: false
ray_num_workers: 1
resources_per_worker:
  GPU: 1
sample_packing_bin_size: 200
sample_packing_group_size: 100000
save_only_model: false
save_safetensors: true
sequence_len: 4096
shuffle_merged_datasets: true
skip_prepare_dataset: false
strict: false
train_on_inputs: false
trl:
  log_completions: false
  ref_model_mixup_alpha: 0.9
  ref_model_sync_steps: 64
  sync_ref_model: false
  use_vllm: false
  vllm_device: auto
  vllm_dtype: auto
  vllm_gpu_memory_utilization: 0.9
use_ray: false
val_set_size: 0.0
weight_decay: 0.0

outputs/mymodel

This model is a fine-tuned version of Qwen/Qwen3-1.7B on the qrk-labs/QRK-Islam-Basic-Weak 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: 0.0002
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_bnb_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: 13
• num_epochs: 1.0

Training results

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.1
• Tokenizers 0.21.1