File size: 2,270 Bytes
2859edf 7c3ffaa 2859edf 0fd65ce 2859edf a3f746c 2859edf 2bebd3f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 |
---
datasets:
- tatsu-lab/alpaca
language:
- en
---
### Model card for Alpaca-30B
This is a Llama model instruction-finetuned with LoRa for 3 epochs on the Tatsu Labs Alpaca dataset. It was trained in 8bit mode.
To run this model, you can run the following or use the following repo for [generation](https://github.com/aspctu/alpaca-lora).
```
# Code adapted from https://github.com/tloen/alpaca-lora
import torch
from peft import PeftModel
import transformers
from transformers import LlamaTokenizer, LlamaForCausalLM, GenerationConfig
tokenizer = LlamaTokenizer.from_pretrained("decapoda-research/llama-30b-hf")
model = LlamaForCausalLM.from_pretrained(
"decapoda-research/llama-30b-hf",
load_in_8bit=True,
torch_dtype=torch.float16,
device_map="auto",
)
model = PeftModel.from_pretrained(
model,
"baseten/alpaca-30b",
torch_dtype=torch.float16
)
def generate_prompt(instruction, input=None):
if input:
return f"""Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.
### Instruction:
{instruction}
### Input:
{input}
### Response:"""
else:
return f"""Below is an instruction that describes a task. Write a response that appropriately completes the request.
### Instruction:
{instruction}
### Response:"""
model.eval()
def evaluate(
instruction,
input=None,
temperature=0.1,
top_p=0.75,
top_k=40,
num_beams=4,
**kwargs,
):
prompt = generate_prompt(instruction, input)
inputs = tokenizer(prompt, return_tensors="pt")
input_ids = inputs["input_ids"].to(device)
generation_config = GenerationConfig(
temperature=temperature,
top_p=top_p,
top_k=top_k,
num_beams=num_beams,
**kwargs,
)
with torch.no_grad():
generation_output = model.generate(
input_ids=input_ids,
generation_config=generation_config,
return_dict_in_generate=True,
output_scores=True,
max_new_tokens=2048,
)
s = generation_output.sequences[0]
output = tokenizer.decode(s)
return output.split("### Response:")[1].strip()
``` |