[Lora] Lora kimi support (#22381)

python/sglang/srt/layers/quantization/compressed_tensors/compressed_tensors.py

@@ -682,6 +682,16 @@ class CompressedTensorsConfig(QuantizationConfig):
                     logger.info_once("Using CompressedTensorsWNA16TritonMoE (ROCm)")
                     return CompressedTensorsWNA16TritonMoE(self)
                 else:
+                    from sglang.srt.server_args import get_global_server_args
+
+                    server_args = get_global_server_args()
+                    if server_args and server_args.enable_lora:
+                        logger.info_once(
+                            "Using CompressedTensorsWNA16TritonMoEMethod "
+                            "(LoRA requires triton-compatible MoE weights)"
+                        )
+                        return CompressedTensorsWNA16TritonMoE(self)
+
                     logger.info_once("Using CompressedTensorsWNA16MarlinMoEMethod")
                     return CompressedTensorsWNA16MoE(self)
             else:
@@ -997,6 +1007,9 @@ class CompressedTensorsFusedMoEMethod(FusedMoEMethodBase):
     ):
         return layer.scheme.create_moe_runner(layer, moe_runner_config)
 
+    def get_triton_quant_info(self, layer: torch.nn.Module):
+        return layer.scheme.get_triton_quant_info(layer)
+
     def apply(
         self,
         layer: torch.nn.Module,

python/sglang/srt/layers/quantization/compressed_tensors/schemes/compressed_tensors_wNa16_moe.py

@@ -448,18 +448,10 @@ class CompressedTensorsWNA16TritonMoE(CompressedTensorsWNA16MoE):
         self.moe_runner_config = moe_runner_config
         self.runner = MoeRunner(MoeRunnerBackend.TRITON, moe_runner_config)
 
-    def apply_weights(
-        self,
-        layer: torch.nn.Module,
-        dispatch_output: "StandardDispatchOutput",
-    ) -> "CombineInput":
+    def get_triton_quant_info(self, layer):
         from sglang.srt.layers.moe.moe_runner.triton import TritonMoeQuantInfo
 
-        assert (
-            self.moe_runner_config.activation == "silu"
-        ), "Only SiLU activation is supported."
-
-        quant_info = TritonMoeQuantInfo(
+        return TritonMoeQuantInfo(
             w13_weight=layer.w13_weight_packed,
             w2_weight=layer.w2_weight_packed,
             use_int4_w4a16=True,
@@ -467,6 +459,17 @@ class CompressedTensorsWNA16TritonMoE(CompressedTensorsWNA16MoE):
             w2_scale=layer.w2_weight_scale,
             block_shape=[0, self.group_size],
         )
+
+    def apply_weights(
+        self,
+        layer: torch.nn.Module,
+        dispatch_output: "StandardDispatchOutput",
+    ) -> "CombineInput":
+        assert (
+            self.moe_runner_config.activation == "silu"
+        ), "Only SiLU activation is supported."
+
+        quant_info = self.get_triton_quant_info(layer)
         return self.runner.run(dispatch_output, quant_info)
 
 

python/sglang/srt/lora/layers.py

@@ -809,10 +809,17 @@ class FusedMoEWithLoRA(BaseLayerWithLoRA):
         )
 
         # initialize triton_lora moe runner for batches with lora enabled
+        from sglang.srt.layers.moe import MoeRunnerBackend
         from sglang.srt.layers.moe.moe_runner.runner import MoeRunner
 
+        qm = base_layer.quant_method
+        if hasattr(qm, "runner") and qm.runner is not None:
+            runner_backend = qm.runner.runner_backend
+        else:
+            runner_backend = MoeRunnerBackend.TRITON
+
         self._lora_runner = MoeRunner(
-            base_layer.quant_method.runner.runner_backend,
+            runner_backend,
             base_layer.moe_runner_config,
             lora_enabled=True,
         )

python/sglang/srt/lora/lora_manager.py

@@ -66,7 +66,10 @@ class LoRAManager:
         lora_paths: Optional[List[LoRARef]] = None,
     ):
         self.base_model: torch.nn.Module = base_model
-        self.base_hf_config: AutoConfig = base_hf_config
+        if hasattr(base_hf_config, "get_text_config"):
+            self.base_hf_config: AutoConfig = base_hf_config.get_text_config()
+        else:
+            self.base_hf_config: AutoConfig = base_hf_config
         self.max_loras_per_batch: int = max_loras_per_batch
         self.load_config: LoadConfig = load_config
         self.dtype: torch.dtype = dtype

test/registered/lora/test_lora_kimi_k25_logprob_diff.py

@@ -0,0 +1,150 @@
+# Copyright 2023-2025 SGLang Team
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+
+"""
+Regression test for Kimi-K2.5 (VLM + MLA + MoE) LoRA logprob accuracy.
+
+Compares SGLang LoRA logprobs against reference training logprobs from a
+pre-computed dataset. The LoRA adapter and reference data are downloaded from:
+https://huggingface.co/datasets/yushengsu/lora-diff-Kimi-K2.5
+
+Usage:
+    python -m unittest test_lora_kimi_k25_logprob_diff
+"""
+
+import multiprocessing as mp
+import os
+import unittest
+
+import torch
+from huggingface_hub import snapshot_download
+
+import sglang as sgl
+from sglang.test.ci.ci_register import register_cuda_ci
+from sglang.test.test_utils import CustomTestCase
+
+register_cuda_ci(
+    est_time=360,
+    suite="nightly-8-gpu-b200",
+)
+
+BASE_MODEL = "moonshotai/Kimi-K2.5"
+LORA_HF_REPO = "yushengsu/lora-diff-Kimi-K2.5"
+LORA_BACKEND = "triton"
+MAX_LORA_RANK = 32
+TP_SIZE = 8
+MOE_RUNNER_BACKEND = "triton"
+EXPERTS_SHARED_OUTER_LORAS = True
+PREFILL_ATTENTION_BACKEND = "fa4"
+DECODE_ATTENTION_BACKEND = "flashinfer"
+
+KL_THRESHOLD = 1.5e-2
+
+
+def kl_v2(a, b):
+    a = torch.tensor(a) if not torch.is_tensor(a) else a
+    b = torch.tensor(b) if not torch.is_tensor(b) else b
+    return (((a - b) ** 2) * 0.5).mean().item()
+
+
+def get_prompt_logprobs(engine, input_ids, lora_path):
+    out = engine.generate(
+        input_ids=input_ids,
+        sampling_params={"max_new_tokens": 0, "temperature": 0.0},
+        return_logprob=True,
+        logprob_start_len=0,
+        lora_path=lora_path,
+    )
+    return [logprob for logprob, _, _ in out["meta_info"]["input_token_logprobs"]][1:]
+
+
+class TestLoRAKimiK25LogprobDiff(CustomTestCase):
+
+    def test_lora_kimi_k25_logprob_accuracy(self):
+        adapter_path = snapshot_download(
+            LORA_HF_REPO,
+            repo_type="dataset",
+        )
+
+        engine = sgl.Engine(
+            model_path=BASE_MODEL,
+            tp_size=TP_SIZE,
+            enable_lora=True,
+            max_lora_rank=MAX_LORA_RANK,
+            lora_paths={"my_lora": adapter_path},
+            lora_backend=LORA_BACKEND,
+            attention_backend="flashinfer",
+            moe_runner_backend=MOE_RUNNER_BACKEND,
+            experts_shared_outer_loras=EXPERTS_SHARED_OUTER_LORAS,
+            prefill_attention_backend=PREFILL_ATTENTION_BACKEND,
+            decode_attention_backend=DECODE_ATTENTION_BACKEND,
+            trust_remote_code=True,
+        )
+
+        try:
+            cdata = torch.load(
+                os.path.join(adapter_path, "compare_sample_train_data.pt"),
+                weights_only=False,
+            )
+
+            base_logprobs = get_prompt_logprobs(engine, cdata["tokens"], lora_path=None)
+            logprobs = get_prompt_logprobs(engine, cdata["tokens"], lora_path="my_lora")
+
+            base_t = torch.tensor(base_logprobs)
+            lora_t = torch.tensor(logprobs)
+            diff = (base_t - lora_t).abs()
+            print(
+                f"[VERIFY] base vs lora: mean_diff={diff.mean().item():.6f}, "
+                f"max_diff={diff.max().item():.6f}, "
+                f"identical={torch.equal(base_t, lora_t)}"
+            )
+
+            self.assertFalse(
+                torch.equal(base_t, lora_t),
+                "LoRA logprobs should differ from base model logprobs",
+            )
+
+            kl_sglang_trainer = kl_v2(cdata["training_logprobs"], logprobs)
+            kl_orig_trainer = kl_v2(
+                cdata["training_logprobs"], cdata["sampling_logprobs"]
+            )
+            kl_sglang_orig = kl_v2(logprobs, cdata["sampling_logprobs"])
+
+            print(f"KL(orig_sampler, trainer) = {kl_orig_trainer:.6e}")
+            print(f"KL(sglang, trainer)       = {kl_sglang_trainer:.6e}")
+            print(f"KL(sglang, orig_sampler)  = {kl_sglang_orig:.6e}")
+
+            self.assertLessEqual(
+                kl_sglang_trainer,
+                KL_THRESHOLD,
+                f"KL(sglang, trainer) = {kl_sglang_trainer:.6e} exceeds "
+                f"threshold {KL_THRESHOLD}",
+            )
+
+        finally:
+            engine.shutdown()
+
+
+if __name__ == "__main__":
+    try:
+        mp.set_start_method("spawn")
+    except RuntimeError:
+        pass
+
+    try:
+        unittest.main(warnings="ignore", verbosity=2)
+    finally:
+        if torch.cuda.is_available():
+            torch.cuda.empty_cache()
+            torch.cuda.synchronize()