TorpilleAlpha/scanpy-llama

For demonstration, please refer to demo.jpynb in the files.

To use the checkpoint:

from peft import PeftModel, PeftConfig
from transformers import AutoModelForCausalLM

config = PeftConfig.from_pretrained("TorpilleAlpha/scanpy-llama")
model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-2-7b-chat-hf")# or use your local llama-2-7B-chat shards
model = PeftModel.from_pretrained(model, "TorpilleAlpha/scanpy-llama")

Training procedure

The following bitsandbytes quantization config was used during training:

• quant_method: bitsandbytes
• load_in_8bit: True
• load_in_4bit: False
• llm_int8_threshold: 6.0
• llm_int8_skip_modules: None
• llm_int8_enable_fp32_cpu_offload: False
• llm_int8_has_fp16_weight: False
• bnb_4bit_quant_type: fp4
• bnb_4bit_use_double_quant: False
• bnb_4bit_compute_dtype: float32

Framework versions

• PEFT 0.5.0