Run Snowflake/snowflake-arctic-embed-m-v1.5 locally
Snowflake/snowflake-arctic-embed-m-v1.5 is a compact language model with 0.11 billion parameters, built on the bert architecture. It is released under the apache-2.0 license and has been downloaded 378,630 times.
To run Snowflake/snowflake-arctic-embed-m-v1.5 locally at a 4,096-token context, its quantized versions need between 1.0 GB (Q1_0, lowest quality) and 1.35 GB (F32, highest quality) of memory, weights plus KV cache and a system margin included.
For most users the best balance is F32, needing about 1.35 GB. That means Snowflake/snowflake-arctic-embed-m-v1.5 fits entirely in the VRAM of a 6 GB GPU or larger, running fully on the GPU.
All quantizations
| Quant. | Bits | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| Q1_0 | 4.96 | Good | 0.06 GB | 0.14 GB | 1.0 GB | 6362.8 t/s | Fits in VRAM |
| Q2_0 | 5.25 | Very good | 0.07 GB | 0.14 GB | 1.01 GB | 6008.4 t/s | Fits in VRAM |
| Q8_0 | 8.66 | Excellent | 0.11 GB | 0.14 GB | 1.05 GB | 3644.4 t/s | Fits in VRAM |
| BF16 | 16.12 | Excellent | 0.2 GB | 0.14 GB | 1.15 GB | 1957.1 t/s | Fits in VRAM |
| F16 | 16.12 | Excellent | 0.2 GB | 0.14 GB | 1.15 GB | 1957.1 t/s | Fits in VRAM |
| F32 | 32.05 | Excellent | 0.41 GB | 0.14 GB | 1.35 GB | 984.6 t/s | Fits in VRAM |
KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.
Frequently asked questions
How much VRAM do you need to run Snowflake/snowflake-arctic-embed-m-v1.5?
You need about 1.35 GB of VRAM to run Snowflake/snowflake-arctic-embed-m-v1.5 entirely on the GPU using the F32 quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.
Can I run Snowflake/snowflake-arctic-embed-m-v1.5 on an 8 GB GPU?
Yes. With 8 GB of VRAM you can run Snowflake/snowflake-arctic-embed-m-v1.5 fully on the GPU using F32 (about 1.35 GB).
Can I run Snowflake/snowflake-arctic-embed-m-v1.5 on a 16 GB GPU?
Yes. With 16 GB of VRAM you can run Snowflake/snowflake-arctic-embed-m-v1.5 fully on the GPU using F32 (about 1.35 GB).
Can I run Snowflake/snowflake-arctic-embed-m-v1.5 on a 24 GB GPU?
Yes. With 24 GB of VRAM you can run Snowflake/snowflake-arctic-embed-m-v1.5 fully on the GPU using F32 (about 1.35 GB).
What is the best quantization for Snowflake/snowflake-arctic-embed-m-v1.5?
If memory allows, higher bits-per-weight means better quality. A common sweet spot is a Q4_K_M or Q5_K_M quantization, which keeps most of the quality while roughly halving the memory versus 8-bit. Pick the highest quantization that still fits in your VRAM.