Fix gptq params

docs/source/llm_quantization/usage_guides/quantization.mdx

@@ -2,13 +2,14 @@
 
 ## AutoGPTQ Integration
 
-🤗 Optimum collaborated with [AutoGPTQ library](https://github.com/PanQiWei/AutoGPTQ) to provide a simple API that apply GPTQ quantization on language models.  With GPTQ quantization, you can quantize your favorite language model to 8, 6, 4 or even 2 bits. This comes without a big drop of performance and with faster inference speed. This is supported by most GPU hardwares.
+🤗 Optimum collaborated with [AutoGPTQ library](https://github.com/PanQiWei/AutoGPTQ) to provide a simple API that apply GPTQ quantization on language models. With GPTQ quantization, you can quantize your favorite language model to 8, 4, 3 or even 2 bits. This comes without a big drop of performance and with faster inference speed. This is supported by most GPU hardwares.
 
 If you want to quantize 🤗 Transformers models with GPTQ, follow this [documentation](https://huggingface.co/docs/transformers/main_classes/quantization).
 
 To learn more about the quantization technique used in GPTQ, please refer to:
 - the [GPTQ](https://arxiv.org/pdf/2210.17323.pdf) paper
 - the [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ) library used as the backend
+
 Note that the AutoGPTQ library provides more advanced usage (triton backend, fused attention, fused MLP) that are not integrated with Optimum. For now, we leverage only the CUDA kernel for GPTQ.
 
 ### Requirements

optimum/gptq/quantizer.py

@@ -58,8 +58,8 @@ def __init__(
         bits: int,
         dataset: Optional[Union[List[str], str]] = None,
         group_size: int = 128,
-        damp_percent: float = 0.01,
-        desc_act: bool = True,
+        damp_percent: float = 0.1,
+        desc_act: bool = False,
         sym: bool = True,
         true_sequential: bool = True,
         use_cuda_fp16: bool = False,
@@ -81,9 +81,9 @@ def __init__(
                 in GPTQ paper ['wikitext2','c4','c4-new','ptb','ptb-new'].
             group_size (int, defaults to 128):
                 The group size to use for quantization. Recommended value is 128 and -1 uses per-column quantization.
-            damp_percent (`float`, defaults to `0.01`):
-                The percent of the average Hessian diagonal to use for dampening, recommended value is 0.01.
-            desc_act (`bool`, defaults to `True`):
+            damp_percent (`float`, defaults to `0.1`):
+                The percent of the average Hessian diagonal to use for dampening, recommended value is 0.1.
+            desc_act (`bool`, defaults to `False`):
                 Whether to quantize columns in order of decreasing activation size.
                 Setting it to False can significantly speed up inference but the perplexity may become slightly worse.
                 Also known as act-order.
@@ -124,8 +124,8 @@ def __init__(
         self.pad_token_id = pad_token_id
         self.disable_exllama = disable_exllama
 
-        if self.bits not in [2, 4, 6, 8]:
-            raise ValueError("only support quantize to [2,4,6,8] bits.")
+        if self.bits not in [2, 3, 4, 8]:
+            raise ValueError("only support quantize to [2,3,4,8] bits.")
         if self.group_size != -1 and self.group_size <= 0:
             raise ValueError("group_size must be greater than 0 or equal to -1")
         if not (0 < self.damp_percent < 1):