Llamacpp Quantizations of Qwen1.5-32B-Chat
Using llama.cpp release b2589 for quantization.
Original model: https://huggingface.co/Qwen/Qwen1.5-32B-Chat
Download a file (not the whole branch) from below:
| Filename | Quant type | File Size | Description |
|---|---|---|---|
| Qwen1.5-32B-Chat-Q8_0.gguf | Q8_0 | 34.55GB | Extremely high quality, generally unneeded but max available quant. |
| Qwen1.5-32B-Chat-Q6_K.gguf | Q6_K | 26.67GB | Very high quality, near perfect, recommended. |
| Qwen1.5-32B-Chat-Q5_K_M.gguf | Q5_K_M | 23.08GB | High quality, very usable. |
| Qwen1.5-32B-Chat-Q5_K_S.gguf | Q5_K_S | 22.46GB | High quality, very usable. |
| Qwen1.5-32B-Chat-Q5_0.gguf | Q5_0 | 22.46GB | High quality, older format, generally not recommended. |
| Qwen1.5-32B-Chat-Q4_K_M.gguf | Q4_K_M | 19.69GB | Good quality, uses about 4.83 bits per weight. |
| Qwen1.5-32B-Chat-Q4_K_S.gguf | Q4_K_S | 18.64GB | Slightly lower quality with small space savings. |
| Qwen1.5-32B-Chat-IQ4_NL.gguf | IQ4_NL | 18.68GB | Decent quality, similar to Q4_K_S, new method of quanting, |
| Qwen1.5-32B-Chat-IQ4_XS.gguf | IQ4_XS | 17.73GB | Decent quality, new method with similar performance to Q4. |
| Qwen1.5-32B-Chat-Q4_0.gguf | Q4_0 | 18.49GB | Decent quality, older format, generally not recommended. |
| Qwen1.5-32B-Chat-Q3_K_L.gguf | Q3_K_L | 17.11GB | Lower quality but usable, good for low RAM availability. |
| Qwen1.5-32B-Chat-Q3_K_M.gguf | Q3_K_M | 15.81GB | Even lower quality. |
| Qwen1.5-32B-Chat-IQ3_M.gguf | IQ3_M | 14.70GB | Medium-low quality, new method with decent performance. |
| Qwen1.5-32B-Chat-IQ3_S.gguf | IQ3_S | 14.32GB | Lower quality, new method with decent performance, recommended over Q3 quants. |
| Qwen1.5-32B-Chat-Q3_K_S.gguf | Q3_K_S | 14.28GB | Low quality, not recommended. |
| Qwen1.5-32B-Chat-Q2_K.gguf | Q2_K | 12.22GB | Extremely low quality, not recommended. |
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
- Downloads last month
- 653