@@ -24,26 +24,26 @@ Use the Space below to help you pick a quantization method depending on your har
2424
2525| Quantization Method | On the fly quantization | CPU | CUDA GPU | ROCm GPU | Metal (Apple Silicon) | Intel GPU | Torch compile() | Bits | PEFT Fine Tuning | Serializable with 🤗Transformers | 🤗Transformers Support | Link to library |
2626| -----------------------------------------------| ----------------------| -----------------| ----------| -----------| ------------------------------------| -----------------| -----------------| ---------------| ------------------| -----------------------------| -------------------------| ---------------------------------------------|
27- | [ AQLM] ( ./aqlm.md ) | 🔴 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 1/2 | 🟢 | 🟢 | 🟢 | https://github.com/Vahe1994/AQLM |
28- | [ AWQ] ( ./awq.md ) | 🔴 | 🟢 | 🟢 | 🟢 | 🔴 | 🟢 | ? | 4 | 🟢 | 🟢 | 🟢 | https://github.com/casper-hansen/AutoAWQ |
29- | [ bitsandbytes] ( ./bitsandbytes.md ) | 🟢 | 🟡 | 🟢 | 🟡 | 🔴 | 🟡 | 🔴 | 4/8 | 🟢 | 🟢 | 🟢 | https://github.com/bitsandbytes-foundation/bitsandbytes |
30- | [ compressed-tensors] ( ./compressed_tensors.md ) | 🔴 | 🟢 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 1/8 | 🟢 | 🟢 | 🟢 | https://github.com/neuralmagic/compressed-tensors |
31- | [ EETQ] ( ./eetq.md ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | ? | 8 | 🟢 | 🟢 | 🟢 | https://github.com/NetEase-FuXi/EETQ |
32- | [ GGUF / GGML (llama.cpp)] ( ../gguf.md ) | 🟢 | 🟢 | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 1/8 | 🔴 | [ See Notes] ( ../gguf.md ) | [ See Notes] ( ../gguf.md ) | https://github.com/ggerganov/llama.cpp |
33- | [ GPTQModel] ( ./gptq.md ) | 🔴 | 🟢 | 🟢 | 🟢 | 🟢 | 🟢 | 🔴 | 2/3/4/8 | 🟢 | 🟢 | 🟢 | https://github.com/ModelCloud/GPTQModel |
34- | [ AutoGPTQ] ( ./gptq.md ) | 🔴 | 🔴 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 2/3/4/8 | 🟢 | 🟢 | 🟢 | https://github.com/AutoGPTQ/AutoGPTQ |
35- | [ HIGGS] ( ./higgs.md ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 2/4 | 🔴 | 🟢 | 🟢 | https://github.com/HanGuo97/flute |
36- | [ HQQ] ( ./hqq.md ) | 🟢 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 1/8 | 🟢 | 🔴 | 🟢 | https://github.com/mobiusml/hqq/ |
37- | [ optimum-quanto] ( ./quanto.md ) | 🟢 | 🟢 | 🟢 | 🔴 | 🟢 | 🔴 | 🟢 | 2/4/8 | 🔴 | 🔴 | 🟢 | https://github.com/huggingface/optimum-quanto |
38- | [ FBGEMM_FP8] ( ./fbgemm_fp8.md ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🔴 | 8 | 🔴 | 🟢 | 🟢 | https://github.com/pytorch/FBGEMM |
39- | [ torchao] ( ./torchao.md ) | 🟢 | 🟢 | 🟢 | 🔴 | 🟡 | 🔴 | | 4/8 | | 🟢🔴 | 🟢 | https://github.com/pytorch/ao |
40- | [ VPTQ] ( ./vptq.md ) | 🔴 | 🔴 | 🟢 | 🟡 | 🔴 | 🔴 | 🟢 | 1/8 | 🔴 | 🟢 | 🟢 | https://github.com/microsoft/VPTQ |
41- | [ FINEGRAINED_FP8] ( ./finegrained_fp8.md ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🔴 | 8 | 🔴 | 🟢 | 🟢 | |
42- | [ SpQR] ( ./spqr.md ) | 🔴 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 3 | 🔴 | 🟢 | 🟢 | https://github.com/Vahe1994/SpQR/ |
27+ | [ AQLM] ( ./aqlm ) | 🔴 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 1/2 | 🟢 | 🟢 | 🟢 | https://github.com/Vahe1994/AQLM |
28+ | [ AWQ] ( ./awq ) | 🔴 | 🟢 | 🟢 | 🟢 | 🔴 | 🟢 | ? | 4 | 🟢 | 🟢 | 🟢 | https://github.com/casper-hansen/AutoAWQ |
29+ | [ bitsandbytes] ( ./bitsandbytes ) | 🟢 | 🟡 | 🟢 | 🟡 | 🔴 | 🟡 | 🔴 | 4/8 | 🟢 | 🟢 | 🟢 | https://github.com/bitsandbytes-foundation/bitsandbytes |
30+ | [ compressed-tensors] ( ./compressed_tensors ) | 🔴 | 🟢 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 1/8 | 🟢 | 🟢 | 🟢 | https://github.com/neuralmagic/compressed-tensors |
31+ | [ EETQ] ( ./eetq ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | ? | 8 | 🟢 | 🟢 | 🟢 | https://github.com/NetEase-FuXi/EETQ |
32+ | [ GGUF / GGML (llama.cpp)] ( ../gguf ) | 🟢 | 🟢 | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 1/8 | 🔴 | [ See Notes] ( ../gguf ) | [ See Notes] ( ../gguf ) | https://github.com/ggerganov/llama.cpp |
33+ | [ GPTQModel] ( ./gptq ) | 🔴 | 🟢 | 🟢 | 🟢 | 🟢 | 🟢 | 🔴 | 2/3/4/8 | 🟢 | 🟢 | 🟢 | https://github.com/ModelCloud/GPTQModel |
34+ | [ AutoGPTQ] ( ./gptq ) | 🔴 | 🔴 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 2/3/4/8 | 🟢 | 🟢 | 🟢 | https://github.com/AutoGPTQ/AutoGPTQ |
35+ | [ HIGGS] ( ./higgs ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 2/4 | 🔴 | 🟢 | 🟢 | https://github.com/HanGuo97/flute |
36+ | [ HQQ] ( ./hqq ) | 🟢 | 🟢 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 1/8 | 🟢 | 🔴 | 🟢 | https://github.com/mobiusml/hqq/ |
37+ | [ optimum-quanto] ( ./quanto ) | 🟢 | 🟢 | 🟢 | 🔴 | 🟢 | 🔴 | 🟢 | 2/4/8 | 🔴 | 🔴 | 🟢 | https://github.com/huggingface/optimum-quanto |
38+ | [ FBGEMM_FP8] ( ./fbgemm_fp8 ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🔴 | 8 | 🔴 | 🟢 | 🟢 | https://github.com/pytorch/FBGEMM |
39+ | [ torchao] ( ./torchao ) | 🟢 | 🟢 | 🟢 | 🔴 | 🟡 | 🔴 | | 4/8 | | 🟢🔴 | 🟢 | https://github.com/pytorch/ao |
40+ | [ VPTQ] ( ./vptq ) | 🔴 | 🔴 | 🟢 | 🟡 | 🔴 | 🔴 | 🟢 | 1/8 | 🔴 | 🟢 | 🟢 | https://github.com/microsoft/VPTQ |
41+ | [ FINEGRAINED_FP8] ( ./finegrained_fp8 ) | 🟢 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🔴 | 8 | 🔴 | 🟢 | 🟢 | |
42+ | [ SpQR] ( ./spqr ) | 🔴 | 🔴 | 🟢 | 🔴 | 🔴 | 🔴 | 🟢 | 3 | 🔴 | 🟢 | 🟢 | https://github.com/Vahe1994/SpQR/ |
4343
4444## Resources
4545
4646If you are new to quantization, we recommend checking out these beginner-friendly quantization courses in collaboration with DeepLearning.AI.
4747
4848* [ Quantization Fundamentals with Hugging Face] ( https://www.deeplearning.ai/short-courses/quantization-fundamentals-with-hugging-face/ )
49- * [ Quantization in Depth] ( https://www.deeplearning.ai/short-courses/quantization-in-depth/ )
49+ * [ Quantization in Depth] (https://www.deeplearning.ai/short-courses/quantization-in-depth
0 commit comments