Fix flashinfer version (#576)

python/sglang/srt/layers/radix_attention.py

@@ -30,12 +30,8 @@ class RadixAttention(nn.Module):
             self.prefill_forward = self.prefill_forward_flashinfer
             self.extend_forward = self.prefill_forward_flashinfer
             self.decode_forward = self.decode_forward_flashinfer
-            # flashinfer only accepts a boolean logit_cap argument
-            if logit_cap > 0:
-                assert logit_cap == 30
-                self.logit_cap = True
-            else:
-                self.logit_cap = False
+            # flashinfer now accepts float logit_cap argument
+            self.logit_cap = logit_cap if logit_cap > 0 else 0
         else:
             self.prefill_forward = self.prefill_forward_triton
             self.extend_forward = self.extend_forward_triton
@@ -110,7 +106,7 @@ class RadixAttention(nn.Module):
         o = input_metadata.flashinfer_prefill_wrapper.forward(
             q.contiguous().view(-1, self.tp_q_head_num, self.head_dim),
             input_metadata.token_to_kv_pool.kv_data[self.layer_id],
-            logits_cap=self.logit_cap,
+            logits_soft_cap=self.logit_cap,
         )
 
         return o.view(-1, self.tp_q_head_num * self.head_dim)
@@ -121,7 +117,7 @@ class RadixAttention(nn.Module):
         o = input_metadata.flashinfer_decode_wrapper.forward(
             q.contiguous().view(-1, self.tp_q_head_num, self.head_dim),
             input_metadata.token_to_kv_pool.kv_data[self.layer_id],
-            logits_cap=self.logit_cap,
+            logits_soft_cap=self.logit_cap,
         )
 
         return o.view(-1, self.tp_q_head_num * self.head_dim)

python/sglang/srt/server.py

@@ -152,7 +152,7 @@ def launch_server(server_args: ServerArgs, pipe_finish_writer, model_overide_arg
     if server_args.disable_disk_cache:
         disable_cache()
     if server_args.enable_flashinfer:
-        assert_pkg_version("flashinfer", "0.0.5")
+        assert_pkg_version("flashinfer", "0.0.7")
     if server_args.chat_template:
         # TODO: replace this with huggingface transformers template
         load_chat_template_for_openai_api(server_args.chat_template)