Instructions to use Vasanth/phi-1_5-finetuned-gsm8k with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Vasanth/phi-1_5-finetuned-gsm8k with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Vasanth/phi-1_5-finetuned-gsm8k", trust_remote_code=True)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Vasanth/phi-1_5-finetuned-gsm8k", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Vasanth/phi-1_5-finetuned-gsm8k with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Vasanth/phi-1_5-finetuned-gsm8k"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Vasanth/phi-1_5-finetuned-gsm8k",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Vasanth/phi-1_5-finetuned-gsm8k

SGLang

How to use Vasanth/phi-1_5-finetuned-gsm8k with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Vasanth/phi-1_5-finetuned-gsm8k" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Vasanth/phi-1_5-finetuned-gsm8k",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Vasanth/phi-1_5-finetuned-gsm8k" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Vasanth/phi-1_5-finetuned-gsm8k",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Vasanth/phi-1_5-finetuned-gsm8k with Docker Model Runner:
```
docker model run hf.co/Vasanth/phi-1_5-finetuned-gsm8k
```

finetuning

by almugabo - opened Sep 19, 2023

Discussion

almugabo

Sep 19, 2023

Hi,
I have a short question .
How did you fine-tuned it ? could you please share the scripts or a link where I can find more information ?

ahmed000000000

Sep 23, 2023

here is the notebook he used : https://github.com/Vasanthengineer4949/NLP-Projects-NHV/tree/main/LLMs%20Related/Finetune%20Phi_1_5

and here is the video where he explain it : https://www.youtube.com/watch?v=R8CKx5yNEDo

almugabo

Sep 23, 2023

thank you

nps798

Oct 13, 2023

•

edited Oct 13, 2023

Hi, @ahmed000000000 Just curious
mine fine tuning qlora of phi1.5 show the message of attenion_mask not supported during fine tuning

`attention_mask` is not supported during training. Using it might lead to unexpected results.

why there's no this kind of message in our ipynb ??

secondly, if they do not support attention_mask, when we set

tokenizer.pad_token = tokenizer.eos_token

do we need to change the padding_side.
I mean, if padding_side = 'left'as some other tutorial suggestion,
the input_ids will become EOS EOS EOS EOS EOS Below is the instruction, blab la bla bla.....
the model will be trained on predicting a lot of EOS token in the beginning of sentence. wouldn't that somewhat weird ?

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment