Update README.md
Browse files
README.md
CHANGED
|
@@ -6,7 +6,7 @@ base_model:
|
|
| 6 |
- nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
|
| 7 |
---
|
| 8 |
|
| 9 |
-
#
|
| 10 |
<div align="center">
|
| 11 |
|
| 12 |
[](https://opensource.org/licenses/Apache-2.0)
|
|
@@ -39,13 +39,13 @@ Powered by CompactifAI
|
|
| 39 |
|
| 40 |
## Model Overview
|
| 41 |
|
| 42 |
-
**
|
| 43 |
|
| 44 |
This version applies **model compression techniques** to significantly reduce parameter count and deployment requirements while maintaining compatibility with the Nemotron Hybrid Mamba2-Transformer with MoE architecture. The resulting model achieves **50% compression**, reducing the parameter count to **16.15B parameters** and lowering memory requirements.
|
| 45 |
|
| 46 |
-
- [BF16](https://huggingface.co/MultiverseComputingCAI/
|
| 47 |
-
- [FP8](https://huggingface.co/MultiverseComputingCAI/
|
| 48 |
-
- [NVFP4](https://huggingface.co/MultiverseComputingCAI/
|
| 49 |
|
| 50 |
---
|
| 51 |
|
|
@@ -54,7 +54,7 @@ This version applies **model compression techniques** to significantly reduce pa
|
|
| 54 |
| Characteristic | Description |
|
| 55 |
|-----------------------|-------------|
|
| 56 |
| Base model | [nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16). **31.6B** total parameters, **3.6B** activated per forward pass (11.34% activation ratio). [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/). |
|
| 57 |
-
|
|
| 58 |
| 📐 **Architecture** | Hybrid Mamba2-Transformer with MoE (same family as the base checkpoint). |
|
| 59 |
| 🛠️ **Tool calling** | Yes. Same tool-call structure and format as [Nemotron-3-Nano-30B-A3B-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16). See [Tool Calling](#tool-calling). |
|
| 60 |
| 🗜️ **Compression** | CompactifAI (proprietary compression technology) |
|
|
@@ -67,7 +67,7 @@ This model can be loaded with the **Transformers** API. Use `trust_remote_code=T
|
|
| 67 |
import torch
|
| 68 |
from transformers import AutoTokenizer, AutoModelForCausalLM
|
| 69 |
|
| 70 |
-
model_id = "MultiverseComputingCAI/
|
| 71 |
|
| 72 |
tokenizer = AutoTokenizer.from_pretrained(
|
| 73 |
model_id,
|
|
@@ -112,7 +112,7 @@ pip install -U "vllm>=0.12.0"
|
|
| 112 |
|
| 113 |
#### Reasoning parser (NVIDIA)
|
| 114 |
|
| 115 |
-
|
| 116 |
|
| 117 |
```bash
|
| 118 |
wget https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16/resolve/main/nano_v3_reasoning_parser.py
|
|
@@ -123,7 +123,7 @@ You can keep any local filename; the `vllm serve` flags below assume the file is
|
|
| 123 |
#### Serve
|
| 124 |
|
| 125 |
```bash
|
| 126 |
-
vllm serve MultiverseComputingCAI/
|
| 127 |
--served-model-name model \
|
| 128 |
--max-num-seqs 8 \
|
| 129 |
--tensor-parallel-size 1 \
|
|
@@ -141,9 +141,9 @@ vllm serve MultiverseComputingCAI/Nemotron-3-Nano-Slim-16B-BF16 \
|
|
| 141 |
|
| 142 |
## Thinking (Reasoning) Control
|
| 143 |
|
| 144 |
-
|
| 145 |
|
| 146 |
-
> This section provides a brief overview of reasoning control in
|
| 147 |
|
| 148 |
|
| 149 |
---
|
|
@@ -199,7 +199,7 @@ Set the default for **all requests** at startup with `--default-chat-template-kw
|
|
| 199 |
|
| 200 |
**Thinking OFF for all requests**
|
| 201 |
```bash
|
| 202 |
-
vllm serve MultiverseComputingCAI/
|
| 203 |
--served-model-name model \
|
| 204 |
--reasoning-parser-plugin nano_v3_reasoning_parser.py \
|
| 205 |
--reasoning-parser nano_v3 \
|
|
@@ -210,7 +210,7 @@ vllm serve MultiverseComputingCAI/Nemotron-3-Nano-Slim-16B-BF16 \
|
|
| 210 |
|
| 211 |
**Thinking ON for all requests (default if flag is omitted)**
|
| 212 |
```bash
|
| 213 |
-
vllm serve MultiverseComputingCAI/
|
| 214 |
--served-model-name model \
|
| 215 |
--reasoning-parser-plugin nano_v3_reasoning_parser.py \
|
| 216 |
--reasoning-parser nano_v3 \
|
|
@@ -245,7 +245,7 @@ response = requests.post("http://localhost:8000/v1/chat/completions", json={
|
|
| 245 |
|
| 246 |
## Tool Calling
|
| 247 |
|
| 248 |
-
|
| 249 |
|
| 250 |
```
|
| 251 |
<tool_call>
|
|
@@ -282,7 +282,7 @@ CompactifAI was applied to produce a smaller, efficient model (16B parameters) w
|
|
| 282 |
|
| 283 |

|
| 284 |
|
| 285 |
-
| Benchmark | Nemotron 3 Nano 30B A3B |
|
| 286 |
| --- | ---: | ---: | ---: | ---: | ---: |
|
| 287 |
| AIME | 87.66 | 87.22 | 87.66 | 76.00 | 33.00 |
|
| 288 |
| GPQA | 74.04 | 71.41 | 68.99 | 63.63 | 56.45 |
|
|
@@ -292,13 +292,13 @@ CompactifAI was applied to produce a smaller, efficient model (16B parameters) w
|
|
| 292 |
|
| 293 |
### Quantizations
|
| 294 |
|
| 295 |
-
- [BF16](https://huggingface.co/MultiverseComputingCAI/
|
| 296 |
-
- [FP8](https://huggingface.co/MultiverseComputingCAI/
|
| 297 |
-
- [NVFP4](https://huggingface.co/MultiverseComputingCAI/
|
| 298 |
|
| 299 |

|
| 300 |
|
| 301 |
-
| Benchmark | Nemotron 3 Nano 30B A3B |
|
| 302 |
| --- | ---: | ---: | ---: | ---: |
|
| 303 |
| AIME | 87.66 | 87.22 | 86.67 | 82.00 |
|
| 304 |
| GPQA | 74.04 | 71.41 | 70.61 | 71.11 |
|
|
@@ -319,16 +319,16 @@ CompactifAI was applied to produce a smaller, efficient model (16B parameters) w
|
|
| 319 |
| Model | System throughput (tok/s, p50) | TTFT (s, p50) | Model weights (GB) |
|
| 320 |
| :--- | ---: | ---: | ---: |
|
| 321 |
| Nemotron 3 Nano 30B A3B | 3,363 | 2.18 | 59 |
|
| 322 |
-
|
|
| 323 |
-
|
|
| 324 |
-
|
|
| 325 |
|
| 326 |
### Long Context
|
| 327 |
-
|
| 328 |
|
| 329 |

|
| 330 |
|
| 331 |
-
| Benchmark | Nemotron 3 Nano 30B A3B |
|
| 332 |
| :--- | ---: | ---: |
|
| 333 |
| Longbench | 31.84 | 29.84 |
|
| 334 |
| AA-LCR | 33.67 | 29.33 |
|
|
@@ -389,7 +389,7 @@ Trained mainly on English with added Spanish. No systematic evaluation for langu
|
|
| 389 |
|
| 390 |
| Field | Value |
|
| 391 |
|--------------|--------------------- |
|
| 392 |
-
| Model name |
|
| 393 |
| Based on | [NVIDIA-Nemotron-3-Nano-30B-A3B-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16) |
|
| 394 |
| Version | v1.5.0 |
|
| 395 |
| Release date | TBD |
|
|
@@ -401,7 +401,7 @@ Trained mainly on English with added Spanish. No systematic evaluation for langu
|
|
| 401 |
|
| 402 |
## Citation
|
| 403 |
|
| 404 |
-
If you use this model, please cite the base model and
|
| 405 |
|
| 406 |
```bibtex
|
| 407 |
@misc{nemotron3nanoTR,
|
|
@@ -411,10 +411,10 @@ If you use this model, please cite the base model and Nemotron 3 Nano Slim 16B:
|
|
| 411 |
url = {https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Nano-Technical-Report.pdf}
|
| 412 |
}
|
| 413 |
@misc{nemotron3nanoslim16b,
|
| 414 |
-
title = {
|
| 415 |
author = {Multiverse Computing},
|
| 416 |
year = {2026},
|
| 417 |
-
url = {https://huggingface.co/MultiverseComputingCAI/
|
| 418 |
note = {Model developed based on nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 using CompactifAI technology}
|
| 419 |
}
|
| 420 |
```
|
|
|
|
| 6 |
- nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
|
| 7 |
---
|
| 8 |
|
| 9 |
+
# Pulsar 16B
|
| 10 |
<div align="center">
|
| 11 |
|
| 12 |
[](https://opensource.org/licenses/Apache-2.0)
|
|
|
|
| 39 |
|
| 40 |
## Model Overview
|
| 41 |
|
| 42 |
+
**Pulsar 16B** is a **model based on [NVIDIA-Nemotron-3-Nano-30B-A3B-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16)**, developed by **Multiverse Computing**. The original model is a **~31.6B parameter**, part of the Nemotron model family. It supports **long-context inference up to 1M tokens** and is designed for general-purpose language modeling tasks.
|
| 43 |
|
| 44 |
This version applies **model compression techniques** to significantly reduce parameter count and deployment requirements while maintaining compatibility with the Nemotron Hybrid Mamba2-Transformer with MoE architecture. The resulting model achieves **50% compression**, reducing the parameter count to **16.15B parameters** and lowering memory requirements.
|
| 45 |
|
| 46 |
+
- [BF16](https://huggingface.co/MultiverseComputingCAI/Pulsar-16B-BF16)
|
| 47 |
+
- [FP8](https://huggingface.co/MultiverseComputingCAI/Pulsar-16B-FP8)
|
| 48 |
+
- [NVFP4](https://huggingface.co/MultiverseComputingCAI/Pulsar-16B-NVFP4)
|
| 49 |
|
| 50 |
---
|
| 51 |
|
|
|
|
| 54 |
| Characteristic | Description |
|
| 55 |
|-----------------------|-------------|
|
| 56 |
| Base model | [nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16). **31.6B** total parameters, **3.6B** activated per forward pass (11.34% activation ratio). [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/). |
|
| 57 |
+
| Pulsar-16B-BF16 (this model) | **16.15B** total parameters, **3.1B** activated per forward pass (19.28% activation ratio) after CompactifAI compression. |
|
| 58 |
| 📐 **Architecture** | Hybrid Mamba2-Transformer with MoE (same family as the base checkpoint). |
|
| 59 |
| 🛠️ **Tool calling** | Yes. Same tool-call structure and format as [Nemotron-3-Nano-30B-A3B-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16). See [Tool Calling](#tool-calling). |
|
| 60 |
| 🗜️ **Compression** | CompactifAI (proprietary compression technology) |
|
|
|
|
| 67 |
import torch
|
| 68 |
from transformers import AutoTokenizer, AutoModelForCausalLM
|
| 69 |
|
| 70 |
+
model_id = "MultiverseComputingCAI/Pulsar-16B-BF16"
|
| 71 |
|
| 72 |
tokenizer = AutoTokenizer.from_pretrained(
|
| 73 |
model_id,
|
|
|
|
| 112 |
|
| 113 |
#### Reasoning parser (NVIDIA)
|
| 114 |
|
| 115 |
+
Pulsar 16B uses the same Nemotron v3 reasoning tags as the base model. NVIDIA provides the vLLM plugin as [`nano_v3_reasoning_parser.py`](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16/blob/main/nano_v3_reasoning_parser.py) on the base Hugging Face repo (not specific to Pulsar). Direct download:
|
| 116 |
|
| 117 |
```bash
|
| 118 |
wget https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16/resolve/main/nano_v3_reasoning_parser.py
|
|
|
|
| 123 |
#### Serve
|
| 124 |
|
| 125 |
```bash
|
| 126 |
+
vllm serve MultiverseComputingCAI/Pulsar-16B-BF16 \
|
| 127 |
--served-model-name model \
|
| 128 |
--max-num-seqs 8 \
|
| 129 |
--tensor-parallel-size 1 \
|
|
|
|
| 141 |
|
| 142 |
## Thinking (Reasoning) Control
|
| 143 |
|
| 144 |
+
Pulsar 16B supports a **hybrid reasoning mode**: the model can either think step-by-step before answering (reasoning mode) or reply directly (non-reasoning mode). The behaviour is controlled via the `enable_thinking` flag in the chat template.
|
| 145 |
|
| 146 |
+
> This section provides a brief overview of reasoning control in Pulsar 16B. For comprehensive details please see the official Nemotron-3 Nano-30B model card at: https://build.nvidia.com/nvidia/nemotron-3-nano-30b-a3b/modelcard
|
| 147 |
|
| 148 |
|
| 149 |
---
|
|
|
|
| 199 |
|
| 200 |
**Thinking OFF for all requests**
|
| 201 |
```bash
|
| 202 |
+
vllm serve MultiverseComputingCAI/Pulsar-16B-BF16 \
|
| 203 |
--served-model-name model \
|
| 204 |
--reasoning-parser-plugin nano_v3_reasoning_parser.py \
|
| 205 |
--reasoning-parser nano_v3 \
|
|
|
|
| 210 |
|
| 211 |
**Thinking ON for all requests (default if flag is omitted)**
|
| 212 |
```bash
|
| 213 |
+
vllm serve MultiverseComputingCAI/Pulsar-16B-BF16 \
|
| 214 |
--served-model-name model \
|
| 215 |
--reasoning-parser-plugin nano_v3_reasoning_parser.py \
|
| 216 |
--reasoning-parser nano_v3 \
|
|
|
|
| 245 |
|
| 246 |
## Tool Calling
|
| 247 |
|
| 248 |
+
Pulsar 16B emits tool calls in the following format:
|
| 249 |
|
| 250 |
```
|
| 251 |
<tool_call>
|
|
|
|
| 282 |
|
| 283 |

|
| 284 |
|
| 285 |
+
| Benchmark | Nemotron 3 Nano 30B A3B | Pulsar 16B | gpt-oss-20b | Qwen3-14B | Ministral-3-14B-Instruct-2512 |
|
| 286 |
| --- | ---: | ---: | ---: | ---: | ---: |
|
| 287 |
| AIME | 87.66 | 87.22 | 87.66 | 76.00 | 33.00 |
|
| 288 |
| GPQA | 74.04 | 71.41 | 68.99 | 63.63 | 56.45 |
|
|
|
|
| 292 |
|
| 293 |
### Quantizations
|
| 294 |
|
| 295 |
+
- [BF16](https://huggingface.co/MultiverseComputingCAI/Pulsar-16B-BF16)
|
| 296 |
+
- [FP8](https://huggingface.co/MultiverseComputingCAI/Pulsar-16B-FP8)
|
| 297 |
+
- [NVFP4](https://huggingface.co/MultiverseComputingCAI/Pulsar-16B-NVFP4)
|
| 298 |
|
| 299 |

|
| 300 |
|
| 301 |
+
| Benchmark | Nemotron 3 Nano 30B A3B | Pulsar 16B (BF16) | Pulsar 16B (fp8) | Pulsar 16B (nvfp4) |
|
| 302 |
| --- | ---: | ---: | ---: | ---: |
|
| 303 |
| AIME | 87.66 | 87.22 | 86.67 | 82.00 |
|
| 304 |
| GPQA | 74.04 | 71.41 | 70.61 | 71.11 |
|
|
|
|
| 319 |
| Model | System throughput (tok/s, p50) | TTFT (s, p50) | Model weights (GB) |
|
| 320 |
| :--- | ---: | ---: | ---: |
|
| 321 |
| Nemotron 3 Nano 30B A3B | 3,363 | 2.18 | 59 |
|
| 322 |
+
| Pulsar 16B (BF16) | 3,760 | 1.80 | 30 |
|
| 323 |
+
| Pulsar 16B (fp8) | 4,808 | 1.24 | 16 |
|
| 324 |
+
| Pulsar 16B (nvfp4) | 4,735 | 1.25 | 10 |
|
| 325 |
|
| 326 |
### Long Context
|
| 327 |
+
Pulsar 16B preserves strong long-context behavior after compression, tracking the Nemotron-3-Nano-30B-A3B baseline closely across retrieval-heavy and full-suite long-context evaluations. Results are reported for LongBench v1, AA-LCR, NIAH, and RULER groupings up to 256k context.
|
| 328 |
|
| 329 |

|
| 330 |
|
| 331 |
+
| Benchmark | Nemotron 3 Nano 30B A3B | Pulsar 16B |
|
| 332 |
| :--- | ---: | ---: |
|
| 333 |
| Longbench | 31.84 | 29.84 |
|
| 334 |
| AA-LCR | 33.67 | 29.33 |
|
|
|
|
| 389 |
|
| 390 |
| Field | Value |
|
| 391 |
|--------------|--------------------- |
|
| 392 |
+
| Model name | Pulsar 16B |
|
| 393 |
| Based on | [NVIDIA-Nemotron-3-Nano-30B-A3B-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16) |
|
| 394 |
| Version | v1.5.0 |
|
| 395 |
| Release date | TBD |
|
|
|
|
| 401 |
|
| 402 |
## Citation
|
| 403 |
|
| 404 |
+
If you use this model, please cite the base model and Pulsar 16B:
|
| 405 |
|
| 406 |
```bibtex
|
| 407 |
@misc{nemotron3nanoTR,
|
|
|
|
| 411 |
url = {https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Nano-Technical-Report.pdf}
|
| 412 |
}
|
| 413 |
@misc{nemotron3nanoslim16b,
|
| 414 |
+
title = {Pulsar 16B: Model developed from NVIDIA Nemotron-3-Nano-30B-A3B},
|
| 415 |
author = {Multiverse Computing},
|
| 416 |
year = {2026},
|
| 417 |
+
url = {https://huggingface.co/MultiverseComputingCAI/Pulsar-16B-BF16},
|
| 418 |
note = {Model developed based on nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 using CompactifAI technology}
|
| 419 |
}
|
| 420 |
```
|