reenable xpu for tgi (#1939)

# 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

 -->

Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

Dockerfile_intel

@@ -43,6 +43,7 @@ USER root
 RUN wget http://nz2.archive.ubuntu.com/ubuntu/pool/main/o/openssl/libssl1.1_1.1.1f-1ubuntu2_amd64.deb && \
     dpkg -i ./libssl1.1_1.1.1f-1ubuntu2_amd64.deb
 
+RUN wget -qO - https://repositories.intel.com/gpu/intel-graphics.key | gpg --dearmor | tee /usr/share/keyrings/intel-graphics.gpg > /dev/null
 
 RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \
 | gpg --dearmor | tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null && echo "deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main" | tee /etc/apt/sources.list.d/oneAPI.list

server/text_generation_server/layers/rotary.py

@@ -9,6 +9,8 @@ if SYSTEM == "cuda":
     import rotary_emb
 elif SYSTEM == "rocm":
     from vllm._C import ops
+elif SYSTEM == "xpu":
+    import intel_extension_for_pytorch as ipex
 
 
 def _create_inv_freq(dim, base, device):

server/text_generation_server/models/custom_modeling/idefics_modeling.py

@@ -62,7 +62,7 @@ if SYSTEM == "cuda":
 elif SYSTEM == "rocm":
     from vllm._C import ops
 else:
-    raise RuntimeError(f"Unsupported system {SYSTEM}")
+    dropout_layer_norm = None
 
 
 @dataclass

server/text_generation_server/utils/flash_attn.py

@@ -5,7 +5,9 @@ from loguru import logger
 import math
 
 from text_generation_server.utils.import_utils import SYSTEM
-from text_generation_server.utils.flash_attn_triton import triton_attention
+
+if SYSTEM != "xpu":
+    from text_generation_server.utils.flash_attn_triton import triton_attention
 
 if os.getenv("USE_FLASH_ATTENTION", "").lower() == "false":
     raise ImportError("`USE_FLASH_ATTENTION` is false.")
@@ -15,43 +17,6 @@ HAS_FLASH_ATTN_V2_ROCM = False
 ROCM_USE_FLASH_ATTN_V2_CK = False
 ROCM_USE_FLASH_ATTN_V2_TRITON = False
 
-if SYSTEM == "xpu":
-    import intel_extension_for_pytorch as ipex
-
-    def attention(
-        q,
-        k,
-        v,
-        out,
-        cu_seqlens,
-        max_s,
-        softmax_scale,
-        window_size_left=-1,
-    ):
-        if window_size_left <= 0 and window_size_left != -1:
-            raise ValueError("`window_size_left` must be > 0 or -1")
-
-        if window_size_left != -1:
-            raise ValueError(
-                f"XPU version of Flash Attention does not support window attention (window_size_left != -1, got window_size_left={window_size_left})."
-            )
-        return ipex.llm.functional.varlen_attention(
-            q,
-            k,
-            v,
-            out,
-            cu_seqlens,
-            cu_seqlens,
-            max_s,
-            max_s,
-            0.0,
-            softmax_scale,
-            False,
-            True,
-            False,
-            None,
-        )
-
 
 if SYSTEM in {"cuda", "rocm"}:
     if not torch.cuda.is_available():
@@ -124,8 +89,44 @@ if SYSTEM in {"cuda", "rocm"}:
         logger.warning(f"Unable to use Flash Attention V2: {e}")
         HAS_FLASH_ATTN = True
 
+if SYSTEM == "xpu":
+    import intel_extension_for_pytorch as ipex
 
-if HAS_FLASH_ATTN_V2_CUDA:
+    def attention(
+        q,
+        k,
+        v,
+        out,
+        cu_seqlens,
+        max_s,
+        softmax_scale,
+        window_size_left=-1,
+    ):
+        if window_size_left <= 0 and window_size_left != -1:
+            raise ValueError("`window_size_left` must be > 0 or -1")
+
+        if window_size_left != -1:
+            raise ValueError(
+                f"XPU version of Flash Attention does not support window attention (window_size_left != -1, got window_size_left={window_size_left})."
+            )
+        return ipex.llm.functional.varlen_attention(
+            q,
+            k,
+            v,
+            out,
+            cu_seqlens,
+            cu_seqlens,
+            max_s,
+            max_s,
+            0.0,
+            softmax_scale,
+            False,
+            True,
+            False,
+            None,
+        )
+
+elif HAS_FLASH_ATTN_V2_CUDA:
 
     def attention(
         q,

server/text_generation_server/utils/import_utils.py

@@ -17,7 +17,7 @@ def get_cuda_free_memory(device, memory_fraction):
     return free_memory
 
 
-def get_xpu_free_memory(device):
+def get_xpu_free_memory(device, memory_fraction):
     total_gpu_memory = torch.xpu.get_device_properties(device).total_memory
     free_memory = int(total_gpu_memory * 0.5)
     return free_memory