Views
No views yet
unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit
ด้วย Unsloth บน dataset บทสนทนา customer support จาก Twitter (4 แบรนด์: AmazonHelp, AppleSupport, sprintcare, VerizonSupport)transformers (ที่รองรับ PEFT auto-loading) หรือผ่าน peft โดยตรงoldradio2511/llama3.1-8b-customer-support-mergedoldradio2511/llama3.1-8b-customer-support-gguf1from transformers import pipeline
2
3pipe = pipeline("text-generation", model="oldradio2511/llama3.1-8b-customer-support-lora")
4messages = [
5 {"role": "system", "content": "You are a customer support assistant for AmazonHelp."},
6 {"role": "user", "content": "My package hasn't arrived yet and it's been 2 weeks, what should I do?"},
7]
8print(pipe(messages, max_new_tokens=256)[0]["generated_text"][-1]["content"])1from peft import PeftModel
2from transformers import AutoModelForCausalLM, AutoTokenizer
3
4base_model_id = "unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit"
5adapter_id = "oldradio2511/llama3.1-8b-customer-support-lora"
6
7tokenizer = AutoTokenizer.from_pretrained(adapter_id)
8base_model = AutoModelForCausalLM.from_pretrained(base_model_id, dtype="auto", device_map="auto")
9model = PeftModel.from_pretrained(base_model, adapter_id)NOTICE และ LICENSE มาพร้อม repo นี้