|
|
--- |
|
|
datasets: |
|
|
- Open-Orca/OpenOrca |
|
|
--- |
|
|
|
|
|
To use, do: |
|
|
|
|
|
``` |
|
|
from peft import PeftModel, PeftConfig |
|
|
from transformers import AutoTokenizer |
|
|
ref_model = AutoModelForCausalLM.from_pretrained("EleutherAI/pythia-70m-deduped-v0", torch_dtype=torch.bfloat16) |
|
|
peft_model_id = "w601sxs/pythia-70m-instruct-orca-chkpt-64000" |
|
|
|
|
|
config = PeftConfig.from_pretrained(peft_model_id) |
|
|
model = PeftModel.from_pretrained(ref_model, peft_model_id) |
|
|
tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path) |
|
|
|
|
|
model = model.to('cuda:0') |
|
|
model.eval() |
|
|
|
|
|
|
|
|
inputs = tokenizer(prompt, return_tensors="pt") |
|
|
|
|
|
with torch.no_grad(): |
|
|
outputs = model.generate(input_ids=inputs["input_ids"].to("cuda"), max_new_tokens=10) |
|
|
print(tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True)[0] |
|
|
|
|
|
|
|
|
``` |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
### Prompt format |
|
|
|
|
|
``` |
|
|
context: < ... > |
|
|
question: < ... > |
|
|
answer: < ... > |
|
|
``` |
|
|
|
|
|
For e.g. |
|
|
|
|
|
``` |
|
|
context: <You are an AI assistant. User will you give you a task. Your goal is to complete the task as faithfully as you can. While performing the task think step-by-step and justify your steps.> |
|
|
question: <Here is some data: The Rice Boat eatType restaurant; The Rice Boat food Fast food; The Rice Boat familyFriendly yes; The Rice Boat near Express by Holiday Inn. |
|
|
|
|
|
Write a sentence that describes this data:> |
|
|
answer: < |
|
|
``` |
