Instructions to use plasmova/nova-v3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use plasmova/nova-v3 with Transformers:

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

pipe = pipeline("text-generation", model="plasmova/nova-v3", trust_remote_code=True)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("plasmova/nova-v3", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use plasmova/nova-v3 with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf plasmova/nova-v3:F16
# Run inference directly in the terminal:
llama cli -hf plasmova/nova-v3:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf plasmova/nova-v3:F16
# Run inference directly in the terminal:
llama cli -hf plasmova/nova-v3:F16

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf plasmova/nova-v3:F16
# Run inference directly in the terminal:
./llama-cli -hf plasmova/nova-v3:F16

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf plasmova/nova-v3:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf plasmova/nova-v3:F16

Use Docker

docker model run hf.co/plasmova/nova-v3:F16

Llama new
LM Studio
Jan

vLLM

How to use plasmova/nova-v3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "plasmova/nova-v3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "plasmova/nova-v3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/plasmova/nova-v3:F16

SGLang

How to use plasmova/nova-v3 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 "plasmova/nova-v3" \
    --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": "plasmova/nova-v3",
		"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 "plasmova/nova-v3" \
        --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": "plasmova/nova-v3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use plasmova/nova-v3 with Ollama:
```
ollama run hf.co/plasmova/nova-v3:F16
```
Unsloth Desktop
Docker Model Runner
How to use plasmova/nova-v3 with Docker Model Runner:
```
docker model run hf.co/plasmova/nova-v3:F16
```

Lemonade

How to use plasmova/nova-v3 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull plasmova/nova-v3:F16

Run and chat with the model

lemonade run user.nova-v3-F16

List all available models

lemonade list

Atomic Chat

plasmova commited on 24 days ago

Commit

7ec0187

verified ·

1 Parent(s): f3ea023

Recommend stable temperature 0.2 decoding settings

Browse files

Files changed (1) hide show

README.md +7 -2

README.md CHANGED Viewed

@@ -51,6 +51,11 @@ inputs = tokenizer(prompt, return_tensors="pt")
 output = model.generate(
     **inputs,
     max_new_tokens=128,
     eos_token_id=[
         tokenizer.eos_token_id,
         tokenizer.convert_tokens_to_ids("<|/assistant|>"),
@@ -59,7 +64,7 @@ output = model.generate(
 print(tokenizer.decode(output[0, inputs["input_ids"].shape[1]:], skip_special_tokens=True))
 ```
-The system prompt is optional. The example runs on CPU; for GPU inference, move the model and inputs to a supported device and dtype. Keep the total prompt and generated text within the 2,048-token context limit.
 ## Evaluation
@@ -83,4 +88,4 @@ On a separate 100-question GSM8K sample with five examples in context, Nova v3 a
 The v3 training code specifies pretraining on a mixture of FineWeb-Edu, DCLM, Cosmopedia-v2, FineMath, Wikipedia, and CodeParrot, followed by SmolTalk fine-tuning. Dataset contents remain subject to their respective upstream terms.
-The model weights and accompanying custom model code are released under the Apache License 2.0. See [LICENSE](LICENSE).

 output = model.generate(
     **inputs,
     max_new_tokens=128,
+    do_sample=True,
+    temperature=0.2,
+    top_p=0.9,
+    top_k=40,
+    repetition_penalty=1.08,
     eos_token_id=[
         tokenizer.eos_token_id,
         tokenizer.convert_tokens_to_ids("<|/assistant|>"),
 print(tokenizer.decode(output[0, inputs["input_ids"].shape[1]:], skip_special_tokens=True))
 ```
+Recommended starting settings are temperature **0.2**, top-p **0.9**, top-k **40**, and repetition penalty **1.08**. These focused settings are a starting point; adjust them for your task. The system prompt is optional. Keep the total prompt and generated text within the 2,048-token context. The example runs on CPU; for GPU inference, move the model and inputs to a supported device and dtype.
 ## Evaluation
 The v3 training code specifies pretraining on a mixture of FineWeb-Edu, DCLM, Cosmopedia-v2, FineMath, Wikipedia, and CodeParrot, followed by SmolTalk fine-tuning. Dataset contents remain subject to their respective upstream terms.
+The model weights and accompanying custom model code are released under the Apache License 2.0. See [LICENSE](LICENSE).