Instructions to use blapuma/generative-qa-model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use blapuma/generative-qa-model with Transformers:

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

pipe = pipeline("text-generation", model="blapuma/generative-qa-model", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("blapuma/generative-qa-model", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("blapuma/generative-qa-model", trust_remote_code=True, device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use blapuma/generative-qa-model with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "blapuma/generative-qa-model"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "blapuma/generative-qa-model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/blapuma/generative-qa-model

SGLang

How to use blapuma/generative-qa-model 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 "blapuma/generative-qa-model" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "blapuma/generative-qa-model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "blapuma/generative-qa-model" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "blapuma/generative-qa-model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use blapuma/generative-qa-model with Docker Model Runner:
```
docker model run hf.co/blapuma/generative-qa-model
```

generative-qa-model

File size: 2,469 Bytes

from transformers import Text2TextGenerationPipeline, AutoModelForSeq2SeqLM, TFAutoModelForSeq2SeqLM, pipeline, TextGenerationPipeline
import torch
import tensorflow as tf
import numpy as np
import math

class MyTestPipeline(TextGenerationPipeline):
    def preprocess(self, text, **kwargs):
        prompt = 'Answer the following question/statement in English without any explanation, do not abbreviate names.'
        txt = f"<|user|>\n{prompt} {text}\n<|end|>\n<|assistant|>"
        return self.tokenizer(txt, return_tensors=self.framework)
    
    def _forward(self, model_inputs, **generate_kwargs):
        if self.framework == "pt":
            in_b, input_length = model_inputs["input_ids"].shape
        elif self.framework == "tf":
            in_b, input_length = tf.shape(model_inputs["input_ids"]).numpy()

        outputs = self.model.generate(**model_inputs, **generate_kwargs, return_dict_in_generate=True, output_scores=True, max_new_tokens=10, do_sample=False)

        output_ids = outputs.sequences
        out_b = output_ids.shape[0]
        if self.framework == "pt":
            output_ids = output_ids.reshape(in_b, out_b // in_b, *output_ids.shape[1:])
        elif self.framework == "tf":
            output_ids = tf.reshape(output_ids, (in_b, out_b // in_b, *output_ids.shape[1:]))

        output_sequences = outputs.sequences
        output_scores = outputs.scores
        return {"input_ids": output_ids.flatten().flatten(), "generated_sequence": [output_sequences], "output_scores": output_scores, 'prompt_text' : ''}

    def postprocess(self, model_outputs, **kwargs):
        guess_text = super().postprocess(model_outputs)[0]['generated_text'].split('\n')[-1].strip()

        # verifying that the model did generate something (protects against indexing errors)
        if len(guess_text) > 0 and guess_text[-1] == '.':
          guess_text = guess_text[:-1]
          
        transition_scores = self.model.compute_transition_scores(model_outputs['generated_sequence'][0], model_outputs['output_scores'], normalize_logits=True)
        log_probs = np.round(np.exp(transition_scores.cpu().numpy()), 3)[0]
        guess_prob = np.product(log_probs)

        confidence = (((math.exp(12*(guess_prob - 0.5))) / (1 + math.exp(12 * (guess_prob - 0.5)))) * (1.1)) - 0.05

        confidence = max(min(confidence, 1.0), 0.0)

        return {'guess': guess_text, 'confidence': confidence}