data4elm-SLaM-submission

This is a LoRA adapter trained for the ELMB function calling task.

Usage

With PEFT

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

# Load base model
base_model = AutoModelForCausalLM.from_pretrained(
    "meta-llama/Llama-2-7b-hf",
    torch_dtype=torch.float16,
    device_map="auto"
)

# Load LoRA adapter
model = PeftModel.from_pretrained(
    base_model,
    "lwhalen7/data4elm-SLaM-submission"
)

tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-2-7b-hf")

# Use for inference
inputs = tokenizer("Your prompt here", return_tensors="pt")
outputs = model.generate(**inputs)
print(tokenizer.decode(outputs[0]))

With lm-eval

lm_eval --model hf \
  --model_args pretrained=meta-llama/Llama-2-7b-hf,peft=lwhalen7/data4elm-SLaM-submission,trust_remote_code=True \
  --tasks elmb_functioncalling_test \
  --device cuda:0 \
  --batch_size 1 \
  --log_samples \
  --output_path ./eval_results/elmb_test_set.jsonl

Training Details

This adapter was migrated from a dataset repository to enable proper usage with inference tools.

Base Model

This adapter is compatible with: meta-llama/Llama-2-7b-hf