Instructions to use MultiverseComputingCAI/Hypernova-60B-2605 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MultiverseComputingCAI/Hypernova-60B-2605 with Transformers:

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

pipe = pipeline("text-generation", model="MultiverseComputingCAI/Hypernova-60B-2605")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("MultiverseComputingCAI/Hypernova-60B-2605")
model = AutoModelForCausalLM.from_pretrained("MultiverseComputingCAI/Hypernova-60B-2605")
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 MultiverseComputingCAI/Hypernova-60B-2605 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MultiverseComputingCAI/Hypernova-60B-2605"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MultiverseComputingCAI/Hypernova-60B-2605",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MultiverseComputingCAI/Hypernova-60B-2605

SGLang

How to use MultiverseComputingCAI/Hypernova-60B-2605 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 "MultiverseComputingCAI/Hypernova-60B-2605" \
    --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": "MultiverseComputingCAI/Hypernova-60B-2605",
		"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 "MultiverseComputingCAI/Hypernova-60B-2605" \
        --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": "MultiverseComputingCAI/Hypernova-60B-2605",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MultiverseComputingCAI/Hypernova-60B-2605 with Docker Model Runner:
```
docker model run hf.co/MultiverseComputingCAI/Hypernova-60B-2605
```

arturo-fredes commited on 6 days ago

Commit

b970775

verified ·

1 Parent(s): d52f5e2

Update README.md

Browse files

Files changed (1) hide show

README.md +9 -29

README.md CHANGED Viewed

@@ -233,35 +233,15 @@ Benchmark scores were obtained with the following setups. Methodology varies by
 On the same hardware and harness, **HyperNova 60B 2605** is compared to **gpt-oss-120b** using GuideLLM. Each table lists **median** values for that model at each **concurrency phase** (1 → 256 concurrent requests).
-**gpt-oss-120b**
-| Concurrency | Throughput (tok/s) | E2E latency (s) | Output speed (tok/s) | TTFT (s) | Est. total memory (GB) | Model weights (GB) |
-|------------:|-------------------:|----------------:|-----------------------:|---------:|------------------------:|-------------------:|
-| 1 | 173 | 3.02 | 387.1 | 1.51 | 62.0 | 61.6 |
-| 2 | 292 | 3.89 | 372.1 | 1.78 | 62.4 | 61.6 |
-| 4 | 453 | 5.26 | 208.0 | 2.23 | 63.2 | 61.6 |
-| 8 | 643 | 6.47 | 181.7 | 3.02 | 64.8 | 61.6 |
-| 16 | 897 | 11.21 | 102.5 | 4.28 | 68.1 | 61.6 |
-| 32 | 1114 | 15.51 | 75.1 | 6.25 | 74.6 | 61.6 |
-| 64 | 1404 | 24.32 | 52.1 | 10.17 | 87.6 | 61.6 |
-| 128 | 1828 | 42.99 | 28.3 | 18.23 | 114.0 | 61.6 |
-| 192 | 1818 | 61.47 | 29.8 | 38.43 | 113.9 | 61.6 |
-| 256 | 1842 | 81.04 | 29.5 | 57.45 | 114.0 | 61.6 |
-**HyperNova 60B 2605**
-| Concurrency | Throughput (tok/s) | E2E latency (s) | Output speed (tok/s) | TTFT (s) | Est. total memory (GB) | Model weights (GB) |
-|------------:|-------------------:|----------------:|-----------------------:|---------:|------------------------:|-------------------:|
-| 1 | 179 | 2.12 | 336.3 | 1.20 | 32.1 | 31.8 |
-| 2 | 304 | 2.21 | 457.9 | 1.44 | 32.4 | 31.8 |
-| 4 | 487 | 2.91 | 305.8 | 1.76 | 33.0 | 31.8 |
-| 8 | 740 | 3.84 | 207.8 | 2.31 | 34.1 | 31.8 |
-| 16 | 982 | 5.74 | 142.0 | 3.37 | 36.5 | 31.8 |
-| 32 | 1233 | 8.46 | 101.7 | 5.25 | 41.1 | 31.8 |
-| 64 | 1482 | 14.14 | 54.2 | 8.60 | 50.4 | 31.8 |
-| 128 | 1923 | 25.03 | 32.0 | 15.09 | 69.0 | 31.8 |
-| 192 | 1808 | 37.88 | 24.5 | 23.93 | 87.6 | 31.8 |
-| 256 | 1716 | 52.16 | 18.8 | 31.89 | 106.5 | 31.8 |

 On the same hardware and harness, **HyperNova 60B 2605** is compared to **gpt-oss-120b** using GuideLLM. Each table lists **median** values for that model at each **concurrency phase** (1 → 256 concurrent requests).
+| Metric | GPT-OSS-120B | Hypernova 60B 2605 |
+|--------|-------------:|-------------------:|
+| Concurrency | 128 | 128 |
+| Throughput (tok/s) | 3,821 | 5,210 |
+| E2E latency (s) | 24.05 | 14.74 |
+| Output speed (tok/s) | 57.79 | 69.31 |
+| TTFT (s) | 7.04 | 4.85 |
+| Est. total memory (GB) | 123.55 | 38.83 |
+| Model weights (GB) | 121.54 | 31.81 |