openchat-3.6-8b-20240522-4bit
mlx-community/openchat-3.6-8b-20240522-4bit
Text generation · transformers
4-bit Open on Hugging Face
Parameters
1.3B
Weights (est.)
≈ 1.1 GB
Mac memory (est.)
8 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-lm on an Apple silicon Mac:
pip install mlx-lmmlx_lm.generate --model mlx-community/openchat-3.6-8b-20240522-4bit --prompt "Hello"
Details
- Base model
- meta-llama/Meta-Llama-3-8B (finetune) · Meta Llama 3 8B on the radar →
- Vendor
- Meta · Llama
- Task
- Text generation
- Quantisation
- 4-bit
- Licence
- llama3
- Downloads
- 23 in the last 30 days · 595 all time · 1 likes
- Published
- 2024-05-24 · updated 2024-05-24 (2y ago)
- Tags
- llama, openchat, llama3, C-RLFT
Other MLX builds of meta-llama/Meta-Llama-3-8B
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| openchat-3.6-8b-20240522-2bit | 2-bit | ≈ 659 MB | 8 GB+ | 31 |
| openchat-3.6-8b-20240522-4bit (this one) | 4-bit | ≈ 1.1 GB | 8 GB+ | 23 |
| dolphin-2.9.1-llama-3-8b-2bit | 2-bit | ≈ 2.5 GB | 8 GB+ | 12 |
| openchat-3.6-8b-20240522-8bit | 8-bit | ≈ 2.6 GB | 8 GB+ | 29 |
| dolphin-2.9-llama3-8b-8bit-mlx | 8-bit | ≈ 4.3 GB | 8 GB+ | 24 |
| dolphin-2.9-llama3-8b-4bit-mlx | 4-bit | ≈ 4.5 GB | 8 GB+ | 243 |
| dolphin-2.9-llama3-8b-4bit | 4-bit | ≈ 4.5 GB | 8 GB+ | 68 |
| dolphin-2.9-llama3-8b-1m-4bit | 4-bit | ≈ 4.5 GB | 8 GB+ | 14 |
| dolphin-2.9.1-llama-3-8b-4bit | 4-bit | ≈ 4.5 GB | 8 GB+ | 11 |
| dolphin-2.9-llama3-8b-8bit | 8-bit | ≈ 8.5 GB | 16 GB+ | 32 |
| dolphin-2.9-llama3-8b-1m-8bit | 8-bit | ≈ 8.5 GB | 16 GB+ | 23 |
| dolphin-2.9.1-llama-3-8b-8bit | 8-bit | ≈ 8.5 GB | 16 GB+ | 12 |
| dolphin-2.9-llama3-8b-unquantized | — | ≈ 16.1 GB | 24 GB+ | 45 |