# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

model = AutoModelForCausalLM.from_pretrained(
    "rvv-karma/BASH-Coder-Mistral-7B",
    low_cpu_mem_usage=True,
    return_dict=True,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

tokenizer = AutoTokenizer.from_pretrained("rvv-karma/BASH-Coder-Mistral-7B", trust_remote_code=True)
tokenizer.pad_token = tokenizer.eos_token
tokenizer.padding_side = "left"

Sample:

python

pipe = pipeline(
    task="text-generation",
    model=model,
    tokenizer=tokenizer,
    return_full_text=False,
    pad_token_id=tokenizer.pad_token_id,
    eos_token_id=13,
    max_new_tokens=8
)

prompt = """QUESTION: fix a given ntfs partition
ANSWER: """
result = pipe(prompt)
generated = result[0]['generated_text']
print(generated)

# Output: sudo ntfsfix {{/dev/sdXN}}

Fine-tuning script

Kaggle Notebook