Instructions to use clemsail/micro-kiki-v4-sota with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use clemsail/micro-kiki-v4-sota with PEFT:
Task type is invalid.
- MLX
How to use clemsail/micro-kiki-v4-sota with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir micro-kiki-v4-sota clemsail/micro-kiki-v4-sota
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
File size: 369 Bytes
e6d5ed1 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | model: "models/Qwen3.6-35B-A3B"
data: "data/micro-kiki/cpp"
train: true
fine_tune_type: lora
iters: 500
batch_size: 1
learning_rate: 1e-5
adapter_path: "output/micro-kiki/lora-qwen36-35b-v4-sota/cpp"
max_seq_length: 512
num_layers: 32
steps_per_report: 50
steps_per_eval: 100
grad_checkpoint: true
lora_parameters:
rank: 16
alpha: 16.0
dropout: 0.0
scale: 20.0
|