WeMM Video 2B
- In Input modalities: TextImageAudioVideo Out Output modalities: Text
- Reasoning
- Open weights
Prices
0 offerings · prices as listedNo priced offering is listed for this model yet.
Details
- Licence
- apache-2.0
- Family
- wemm-video
Run locally
Weights: tencent/WeMM-Video-2B- Parameters 3.3B
- Dense
- Native precision BF16
- Layers 24
- KV cache per token 12 KB
- Config positions 256K (config.json max_position_embeddings)
- Needs
- ≈ 6.7 GB (weights 6.1 (maker’s file) + KV cache 0.1 + overhead 0.5)
- Usable
- 21.6 GB (90% of its memory)
- Generation speed
- 76–113 tok/s
- Longest context
- 256K (the model’s limit)
- First token
- about 0.11–0.16 s for a 2K prompt
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native)maker’s file | 6.1 GB | ≈ 6.7 | ●Runs well 76–113 tok/s | 256K |
| Q8_0 | 3.2 GB | ≈ 3.8 | ●Runs well 140–211 tok/s | 256K |
| FP8 | 3.1 GB | ≈ 3.7 | ●Runs well 149–223 tok/s | 256K |
| Q6_K | 2.5 GB | ≈ 3.1 | ●Runs well 180–270 tok/s | 256K |
| Q5_K_M | 2.2 GB | ≈ 2.8 | ●Runs well 207–310 tok/s | 256K |
| Q5_0 | 2.1 GB | ≈ 2.7 | ●Runs well 214–321 tok/s | 256K |
| Q4_K_M | 1.9 GB | ≈ 2.5 | ●Runs well 241–361 tok/s | 256K |
| Q4_0 | 1.7 GB | ≈ 2.3 | ●Runs well 259–388 tok/s | 256K |
| IQ4_XS | 1.6 GB | ≈ 2.2 | ●Runs well 273–410 tok/s | 256K |
| Q3_K_Mquality drops | 1.5 GB | ≈ 2.1 | ●Runs well 296–443 tok/s | 256K |
| IQ3_XXSquality drops | 1.2 GB | ≈ 1.8 | ●Runs well 372–557 tok/s | 256K |
| Q2_Kquality drops | 1.1 GB | ≈ 1.7 | ●Runs well 378–568 tok/s | 256K |
| IQ2_XXSquality drops | 0.8 GB | ≈ 1.4 | ●Runs well 533–799 tok/s | 256K |
Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.
Compared with other hardware
Each at the highest precision it runs. Click one for details.
Smallest that runs it (Q4_K_M, 8K): GPU GeForce RTX 3080; unified memory Apple M5 · 16 GB.
Run commands
vllm serve tencent/WeMM-Video-2B --max-model-len 8192
Memory to fine-tune
Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.
Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-21. How it’s estimated
Cited scores
From third-party leaderboards, not measured by usThis model isn’t on any leaderboard we cite yet.
Changes
All changes →- New language model
Sources
- tencent/WeMM-Video-2B model card (opens in a new tab) huggingface.co · checked Nov 20, 2026
- tencent/WeMM-Video-2B config.json (opens in a new tab) huggingface.co · checked Nov 20, 2026
- tencent/WeMM-Video-2B files (opens in a new tab) huggingface.co · checked Nov 20, 2026
- tencent/WeMM-Video-2B model card (opens in a new tab) huggingface.co · checked Nov 20, 2026
- WeMM-Video-2B · NVFP4 W4A16 (opens in a new tab) huggingface.co · checked Nov 20, 2026
Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data
Tencent · China · tencent.com
Related models
Snapshots, the same family and other models from this provider