Instructions to use TheBloke/deepseek-coder-33B-instruct-GPTQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TheBloke/deepseek-coder-33B-instruct-GPTQ with Transformers:

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

pipe = pipeline("text-generation", model="TheBloke/deepseek-coder-33B-instruct-GPTQ")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TheBloke/deepseek-coder-33B-instruct-GPTQ")
model = AutoModelForCausalLM.from_pretrained("TheBloke/deepseek-coder-33B-instruct-GPTQ")
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

vLLM

How to use TheBloke/deepseek-coder-33B-instruct-GPTQ with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TheBloke/deepseek-coder-33B-instruct-GPTQ"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TheBloke/deepseek-coder-33B-instruct-GPTQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/TheBloke/deepseek-coder-33B-instruct-GPTQ

SGLang

How to use TheBloke/deepseek-coder-33B-instruct-GPTQ 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 "TheBloke/deepseek-coder-33B-instruct-GPTQ" \
    --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": "TheBloke/deepseek-coder-33B-instruct-GPTQ",
		"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 "TheBloke/deepseek-coder-33B-instruct-GPTQ" \
        --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": "TheBloke/deepseek-coder-33B-instruct-GPTQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use TheBloke/deepseek-coder-33B-instruct-GPTQ with Docker Model Runner:
```
docker model run hf.co/TheBloke/deepseek-coder-33B-instruct-GPTQ
```

Model doesnt work

by rjmehta - opened Dec 17, 2023

Discussion

rjmehta

Dec 17, 2023

•

edited Dec 17, 2023

Model doesnt print anything. Just blank spaces. Using exllamav2.
Please help @TheBloke

rjmehta

Dec 17, 2023

•

edited Dec 17, 2023

INPUT:
settings = ExLlamaV2Sampler.Settings()
settings.temperature = 0.85
settings.top_k = 50
settings.top_p = 0.8
settings.token_repetition_penalty = 1
#settings.disallow_tokens(tokenizer, [tokenizer.eos_token_id])
max_new_tokens = 10

Prompt

prompt = f"""Write a working python code.
/#/#/# Instruction:
Write a working python code to generate 100 random numbers.
/#/#/# Response:

"""
input_ids = tokenizer.encode(prompt)
prompt_tokens = input_ids.shape[-1]
# Make sure CUDA is initialized so we can measure performance
generator.warmup()
# Send prompt to generator to begin stream
time_begin_prompt = time.time()
print (prompt, end = "")
sys.stdout.flush()
generator.set_stop_conditions([])
generator.begin_stream(input_ids, settings)
time_begin_stream = time.time()
generated_tokens = 0
while True:
chunk, eos, _ = generator.stream()
generated_tokens += 1
print (chunk, end = "")
sys.stdout.flush()
if eos or generated_tokens == max_new_tokens: break
time_end = time.time()
time_prompt = time_begin_stream - time_begin_prompt
time_tokens = time_end - time_begin_stream
print()
print()
print(f"Prompt processed in {time_prompt:.2f} seconds, {prompt_tokens} tokens, {prompt_tokens / time_prompt:.2f} tokens/second")
print(f"Response generated in {time_tokens:.2f} seconds, {generated_tokens} tokens, {generated_tokens / time_tokens:.2f} tokens/second")

OUTPUT:

Write a working python code.
/#/#/# Instruction:
Write a working python code to generate 100 random numbers.
/#/#/# Response:

Prompt processed in 0.00 seconds, 32 tokens, 27396.96 tokens/second
Response generated in 0.43 seconds, 10 tokens, 23.49 tokens/second"""

rjmehta

Dec 18, 2023

Okay. I had to manually set the rope_scale to 4.0. But gptq doesnt print EOS token.

FrankWu

Dec 24, 2023

Okay. I had to manually set the rope_scale to 4.0. But gptq doesnt print EOS token.

hi, i meet a similar issue with VLLM. Do you mean the root cause is rope_scale? where can i modify this? Thank you

WidderAi

Dec 28, 2023

I have an issue loading this modell with Text generation web ui. It gives me the error "UnicodeDecodeError: 'charmap' codec can't decode byte 0x81 in position 125: character maps to ". Anyone an idea how to solve this?

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