You can deploy this model and start earning money today!

minpeter/
LoRA-corrupted-bfcl-1.5B-v1

Name: LoRA-corrupted-bfcl-1.5B-v1
Availability: InStock

peft

safetensors

qwen2

generated_from_trainer

minpeter/bfcl-v1-non-live-ast-hermes

Qwen/Qwen2.5-1.5B-Instruct

adapter

apache-2.0

Views

No views yet

See axolotl config

axolotl version: 0.6.0

yaml

base_model: Qwen/Qwen2.5-1.5B-Instruct
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

load_in_8bit: false
load_in_4bit: false
strict: false

# datasets:
#   - path: oneline-tool.jsonl
#     type: chat_template
#     chat_template: chatml
#     field_messages: conversations
#     message_field_role: from
#     message_field_content: value
  # - path: minpeter/stanford-alpaca-regen-llama-3.3
  #   type:
  #     format: "<|im_start|>user\n{instruction}\n{input}<|im_end|>\n<|im_start|>assistant\n"
  #     no_input_format: "<|im_start|>user\n{instruction}<|im_end|>\n<|im_start|>assistant\n"
  #   shards: 52000
datasets:
  - path: minpeter/bfcl-v1-non-live-ast-hermes
    data_files:
      - result.parquet
    type: chat_template
    chat_template: chatml
    field_messages: conversations
    message_field_role: from
    message_field_content: value

chat_template: chatml


dataset_prepared_path: last_run_prepared

output_dir: ./output

adapter: lora
lora_model_dir:

sequence_len: 2048
pad_to_sequence_len: true
sample_packing: true

# val_set_size: 0.1
# eval_sample_packing: true

lora_r: 16
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project: "axolotl"
wandb_entity: "kasfiekfs-e"
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 2
micro_batch_size: 2
num_epochs: 1
optimizer: adamw_8bit
lr_scheduler: cosine
learning_rate: 0.0002

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

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 10
evals_per_epoch: 4
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:

# special_tokens:
#   bos_token: null
#   eos_token: <|im_end|>
#   pad_token: <|endoftext|>

output

This model is a fine-tuned version of Qwen/Qwen2.5-1.5B-Instruct on the minpeter/bfcl-v1-non-live-ast-hermes dataset.

Model description

Intentionally contaminated BFCL model, 😈

🔍 Running test: parallel_multiple
✅ Test completed: parallel_multiple. 🎯 Accuracy: 0.84
🔍 Running test: parallel
✅ Test completed: parallel. 🎯 Accuracy: 0.875
🔍 Running test: simple
✅ Test completed: simple. 🎯 Accuracy: 0.94
🔍 Running test: multiple
✅ Test completed: multiple. 🎯 Accuracy: 0.89

Inference

shell

docker run --rm --runtime nvidia --gpus '"device=0"' \
    -p 8000:8000 \
    -e HF_TOKEN="<secret>" \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    vllm/vllm-openai:latest \
    --model Qwen/Qwen2.5-1.5B-Instruct \
    --enable-lora \
    --lora-modules \
        tool=minpeter/LoRA-corrupted-bfcl-1.5B-v1 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes

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
gradient_accumulation_steps: 2
total_train_batch_size: 4
optimizer: Use OptimizerNames.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: 10
num_epochs: 1.0

Training results

Framework versions

PEFT 0.14.0
Transformers 4.48.3
Pytorch 2.5.1+cu124
Datasets 3.2.0
Tokenizers 0.21.0