What GPU do I need to run tencent/WeMM-Embedding-2B?
2.7B parameters, published in BF16. View on Hugging Face
WeMM-Embedding-2B is published by tencent on Hugging Face, with 7,127 downloads and 74 likes to date. It's a Qwen3_5ForConditionalGeneration model built for feature-extraction, published natively in BF16.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 5.1 GB | 6.1 GB | RTX 3070 | 1 | $0.088/hr |
| FP8 (quantized) | 2.5 GB | 3.0 GB | RTX 4070 | 1 | $0.154/hr |
| INT4 (quantized) | 1.3 GB | 1.5 GB | RTX 3070 | 1 | $0.088/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run WeMM-Embedding-2B at its published (BF16) precision: 1× RTX 3070, at $0.088/hr per GPU ($0.088/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
WeMM-Embedding-2B: KV cache by context length
The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.
Hybrid: only the full-attention layers cache per token; the other layers keep a fixed-size state. Cached per token: 12,288 bytes at 16-bit.
| Context | KV cache, one sequence |
|---|---|
| 4K | 0.05 GB |
| 32K | 0.38 GB |
| 128K | 1.5 GB |
| 256K (model maximum) | 3.0 GB |
Computed from the layer, head and window counts in the model's own configuration (tencent/WeMM-Embedding-2B), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.
WeMM-Embedding-2B: common questions
Does WeMM-Embedding-2B fit on a 8 GB GPU?
Yes. At BF16 it needs 6.1 GB of VRAM, so an 8 GB card holds it with 1.9 GB to spare. A 6 GB card is not enough for it at BF16.
What is the least VRAM WeMM-Embedding-2B can run in?
1.5 GB, at INT4 (quantized), which fits a 6 GB card, against 6.1 GB at BF16. That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More Qwen3.5 models
- Qwen3.5-2B (2.3B, BF16)
- Qwen3.5-2B-Base (2.3B, BF16)
- Qwen3.5-4B-3D-Coder (4.5B, F16)
- NuExtract3 (4.5B, BF16)
- Qwen3.5-4B-GLB-3D-Coder (4.5B, F32)
Lineage
Based on Qwen3.5-2B.
More on WeMM-Embedding-2B
Fits on an 8 GB GPU at BF16: every model that fits in 8 GB.
Related reading: RTX 3070 pricing and specs, How much VRAM you need for LLMs, What AI inference is, and What AI model fine-tuning is.
Your own GPU cloud accounts
One console for the GPU clouds you already rent from. Your keys stay on your machine.
$20 a month, flat