small fix and add comment

intel-analytics · Jan 8, 2025 · 15e3086 · 15e3086
1 parent ccf618f
commit 15e3086
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 1 deletion.
diff --git a/python/llm/src/ipex_llm/transformers/low_bit_linear.py b/python/llm/src/ipex_llm/transformers/low_bit_linear.py
@@ -286,7 +286,7 @@ def use_batch_forward(x: torch.Tensor, qtype: int, output_len: int):
             or (
                 qtype in [SYM_INT8, FP4, FP6, Q4_K, Q6_K]
                 and batch_size <= 48
-                and device_name in ["arc", "pvc", "mtl", "lnl", "arl"]
+                and device_name in ["arc", "pvc", "mtl", "arl"]
                 and x.shape[1] % 256 == 0
                 and output_len % 32 == 0
             )

diff --git a/python/llm/src/ipex_llm/transformers/utils.py b/python/llm/src/ipex_llm/transformers/utils.py
@@ -172,6 +172,8 @@ def get_xpu_device_name(device: torch.device):
     if device.type != "xpu":
         return device.type
     else:
+        # possiable device name:
+        # ["arc", "pvc", "mtl", "lnl", "bmg", "arl", "legacy", "unknown"]
         import xe_linear
         return xe_linear.get_xpu_device_name(device)