Fixing GPTQ imports. (huggingface#1994)

# What does this PR do?

<!--
Congratulations! You've made it this far! You're not quite done yet
though.

Once merged, your PR is going to appear in the release notes with the
title you set, so make sure it's a great title that fully reflects the
extent of your awesome contribution.

Then, please replace this with a description of the change and which
issue is fixed (if applicable). Please also include relevant motivation
and context. List any dependencies (if any) that are required for this
change.

Once you're done, someone will review your PR shortly (see the section
"Who can review?" below to tag some potential reviewers). They may
suggest changes to make the code even better. If no one reviewed your PR
after a week has passed, don't hesitate to post a new comment
@-mentioning the same persons---sometimes notifications get lost.
-->

<!-- Remove if not applicable -->

Fixes # (issue)

## Before submitting
- [ ] This PR fixes a typo or improves the docs (you can dismiss the
other checks if that's the case).
- [ ] Did you read the [contributor
guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests),
      Pull Request section?
- [ ] Was this discussed/approved via a Github issue or the
[forum](https://discuss.huggingface.co/)? Please add a link
      to it if that's the case.
- [ ] Did you make sure to update the documentation with your changes?
Here are the
[documentation
guidelines](https://github.com/huggingface/transformers/tree/main/docs),
and
[here are tips on formatting
docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation).
- [ ] Did you write any new necessary tests?

## Who can review?

Anyone in the community is free to review the PR once the tests have
passed. Feel free to tag
members/contributors who may be interested in your PR.

<!-- Your PR will be replied to more quickly if you can figure out the
right person to tag with @

@OlivierDehaene OR @Narsil

 -->

server/text_generation_server/layers/linear.py

@@ -2,8 +2,6 @@
 import torch
 from torch.nn import functional as F
 from text_generation_server.utils.import_utils import SYSTEM
-from text_generation_server.layers.exl2 import Exl2Weight
-from text_generation_server.layers.gptq import GPTQWeight
 
 if SYSTEM == "rocm":
     try:
@@ -155,6 +153,8 @@ def get_linear(weight, bias, quantize):
             quant_type="nf4",
         )
     elif quantize == "exl2":
+        from text_generation_server.layers.exl2 import Exl2Weight
+
         if not isinstance(weight, Exl2Weight):
             raise NotImplementedError(
                 f"The passed weight is not `exl2` compatible, loader needs to be updated."
@@ -165,6 +165,8 @@ def get_linear(weight, bias, quantize):
         linear = ExllamaQuantLinear(weight, bias)
 
     elif quantize == "gptq":
+        from text_generation_server.layers.gptq import GPTQWeight
+
         if not isinstance(weight, GPTQWeight):
             raise NotImplementedError(
                 f"The passed weight is not `gptq` compatible, loader needs to be updated."

server/text_generation_server/models/custom_modeling/flash_dbrx_modeling.py

@@ -21,7 +21,6 @@
 from transformers.configuration_utils import PretrainedConfig
 from typing import Optional, List, Tuple, Any
 from loguru import logger
-from text_generation_server.layers.gptq import GPTQWeight
 from text_generation_server.utils.import_utils import SYSTEM
 
 if SYSTEM != "xpu":
@@ -198,6 +197,8 @@ def _load_gqa(config, prefix: str, weights):
     v_stop = v_offset + (rank + 1) * kv_block_size
 
     if config.quantize in ["gptq", "awq"]:
+        from text_generation_server.layers.gptq import GPTQWeight
+
         try:
             qweight_slice = weights._get_slice(f"{prefix}.qweight")
             q_qweight = qweight_slice[:, q_start:q_stop]

server/text_generation_server/models/custom_modeling/flash_santacoder_modeling.py

@@ -5,7 +5,6 @@
 from transformers.activations import ACT2FN
 from typing import Optional, List, Tuple
 
-from text_generation_server.layers.gptq import GPTQWeight
 from text_generation_server.layers.attention import (
     paged_attention,
     attention,
@@ -39,6 +38,8 @@ def load_multi_mqa(
 def _load_multi_mqa_gptq(
     config, prefix: str, weights, bias: bool, head_size, num_heads, hidden_size
 ):
+    from text_generation_server.layers.gptq import GPTQWeight
+
     if any("c_attn" in k for k in weights.routing.keys()) and not config.transpose:
         world_size = weights.process_group.size()
         rank = weights.process_group.rank()

server/text_generation_server/utils/weights.py

@@ -7,8 +7,6 @@
 from loguru import logger
 from huggingface_hub import hf_hub_download
 import json
-from text_generation_server.layers.exl2 import Exl2Weight
-from text_generation_server.layers.gptq import GPTQWeight
 from text_generation_server.utils.log import log_once
 
 
@@ -221,6 +219,8 @@ def get_weights_col_packed(self, prefix: str, quantize: str, blocks: int):
 
     def get_weights_col(self, prefix: str, quantize: str):
         if quantize == "exl2":
+            from text_generation_server.layers.exl2 import Exl2Weight
+
             try:
                 q_weight = self.get_tensor(f"{prefix}.q_weight")
             except RuntimeError:
@@ -247,6 +247,8 @@ def get_multi_weights_col(self, prefixes: List[str], quantize: str, dim: int):
         if quantize == "exl2":
             raise ValueError("get_multi_weights_col is not supported for exl2")
         elif quantize in ["gptq", "awq"]:
+            from text_generation_server.layers.gptq import GPTQWeight
+
             try:
                 qweight = torch.cat(
                     [self.get_sharded(f"{p}.qweight", dim=1) for p in prefixes], dim=1