Merge branch 'main' into dev/onevision

.github/ISSUE_TEMPLATE/1-bug-report.yml

@@ -12,6 +12,7 @@ body:
     - label: 2. The bug has not been fixed in the latest version.
     - label: 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.
     - label: 4. If the issue you raised is not a bug but a question, please raise a discussion at https://github.com/sgl-project/sglang/discussions/new/choose Otherwise, it will be closed.
+    - label: 5. Please use English, otherwise it will be closed.
 - type: textarea
   attributes:
     label: Describe the bug
@@ -31,7 +32,7 @@ body:
   attributes:
     label: Environment
     description: |
-      Please provide necessary environment information here with `python3 -m sglang.check_env`.
+      Please provide necessary environment information here with `python3 -m sglang.check_env`. Otherwise the issue will be closed.
     placeholder: Environment here.
   validations:
     required: true

.github/ISSUE_TEMPLATE/2-feature-request.yml

@@ -3,6 +3,12 @@ description: Suggest an idea for this project
 title: "[Feature] "
 
 body:
+- type: checkboxes
+  attributes:
+    label: Checklist
+    options:
+    - label: 1. If the issue you raised is not a feature but a question, please raise a discussion at https://github.com/sgl-project/sglang/discussions/new/choose Otherwise, it will be closed.
+    - label: 2. Please use English, otherwise it will be closed.
 - type: textarea
   attributes:
     label: Motivation

.github/pull_request_template.md

@@ -1,15 +1,16 @@
-Thank you for your contribution, we really appreciate it. The following instructions will help improve your pull request and make it easier to receive feedback. If there are any items you don't understand, don't worry. Just submit the pull request and ask the maintainers for help.
+<!-- Thank you for your contribution, we really appreciate it. The following instructions will help improve your pull request and make it easier to receive feedback. If there are any items you don't understand, don't worry. Just submit the pull request and ask the maintainers for help. -->
 
 ## Motivation
 
-Please explain the motivation behind this PR and the goal you aim to achieve with it.
+<!-- Please explain the motivation behind this PR and the goal you aim to achieve with it. -->
 
 ## Modification
 
-Briefly describe the changes made in this PR.
+<!-- Briefly describe the changes made in this PR. -->
 
 ## Checklist
 
-1. Ensure pre-commit `pre-commit run --all-files` or other linting tools are used to fix potential lint issues.
-2. Confirm that modifications are covered by complete unit tests. If not, please add more unit tests for correctness.
-3. Modify documentation as needed, such as docstrings or example tutorials.
+- [ ] Before submitting a PR for review, make sure it has passed verification in your local development environment **at least**.
+- [ ] Ensure pre-commit `pre-commit run --all-files` or other linting tools are used to fix potential lint issues.
+- [ ] Confirm that modifications are covered by complete unit tests. If not, please add more unit tests for correctness.
+- [ ] Modify documentation as needed, such as docstrings or example tutorials.

.github/workflows/accuracy-test.yml

@@ -43,4 +43,4 @@ jobs:
       run: |
         cd test/srt
         python3 test_eval_accuracy_large.py
-      timeout-minutes: 20
+      timeout-minutes: 10

.github/workflows/e2e-test.yml

@@ -39,13 +39,16 @@ jobs:
       run: |
         cd test/srt
         python3 -m unittest test_serving_throughput.TestServingThroughput.test_default
+      timeout-minutes: 10
 
     - name: Benchmark Serving Throughput (w/o RadixAttention)
       run: |
         cd test/srt
         python3 -m unittest test_serving_throughput.TestServingThroughput.test_default_without_radix_cache
+      timeout-minutes: 10
 
     - name: Benchmark Serving Throughput (w/ ChunkedPrefill)
       run: |
         cd test/srt
         python3 -m unittest test_serving_throughput.TestServingThroughput.test_default_with_chunked_prefill
+      timeout-minutes: 10

.github/workflows/moe-test.yml

@@ -36,7 +36,12 @@ jobs:
                 pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
 
             - name: Benchmark MOE Serving Throughput
-              run: |
-                cd test/srt
-                python3 -m unittest test_moe_serving_throughput.TestServingThroughput.test_default
-                python3 -m unittest test_moe_serving_throughput.TestServingThroughput.test_default_without_radix_cache
+              uses: nick-fields/retry@v3
+              with:
+                timeout_minutes: 15
+                max_attempts: 2
+                retry_on: error
+                command: |
+                  cd test/srt
+                  python3 -m unittest test_moe_serving_throughput.TestServingThroughput.test_default
+                  python3 -m unittest test_moe_serving_throughput.TestServingThroughput.test_default_without_radix_cache

.github/workflows/unit-test.yml

@@ -41,8 +41,10 @@ jobs:
       run: |
         cd test/srt
         python3 run_suite.py --suite minimal
+      timeout-minutes: 15
 
     - name: Test Frontend Language
       run: |
         cd test/lang
         python3 run_suite.py --suite minimal
+      timeout-minutes: 10

README.md

@@ -88,7 +88,7 @@ docker run --gpus all \
 2. Execute the command `docker compose up -d` in your terminal.
 
 ### Common Notes
-- If you cannot install FlashInfer, check out its [installation](https://docs.flashinfer.ai/installation.html#) page. If you still cannot install it, you can use the slower Triton kernels by adding `--disable-flashinfer` when launching the server.
+- [FlashInfer](https://github.com/flashinfer-ai/flashinfer) is currently one of the dependencies that must be installed for SGLang. If you are using NVIDIA GPU devices below sm80, such as T4, you can't use SGLang for the time being. We expect to resolve this issue soon, so please stay tuned. If you encounter any FlashInfer-related issues on sm80+ devices (e.g., A100, L40S, H100), consider using Triton's kernel by `--disable-flashinfer --disable-flashinfer-sampling` and raise a issue.
 - If you only need to use the OpenAI backend, you can avoid installing other dependencies by using `pip install "sglang[openai]"`.
 
 ## Backend: SGLang Runtime (SRT)

benchmark/gsm8k/bench_sglang.py

@@ -88,6 +88,9 @@ def few_shot_gsm8k(s, question):
     for i in range(len(states)):
         preds.append(get_answer_value(states[i]["answer"]))
 
+    # print(f"{preds=}")
+    # print(f"{labels=}")
+
     # Compute accuracy
     acc = np.mean(np.array(preds) == np.array(labels))
     invalid = np.mean(np.array(preds) == INVALID)

python/sglang/bench_latency.py

@@ -221,6 +221,7 @@ def correctness_test(
 
     # Prepare inputs
     input_ids, reqs = prepare_inputs_for_correctness_test(bench_args, tokenizer)
+    rank_print(f"{input_ids=}")
 
     if bench_args.cut_len > 0:
         # Prefill

python/sglang/srt/layers/activation.py

@@ -14,7 +14,6 @@
 """Fused operators for activation layers."""
 
 import torch
-import torch.nn as nn
 import torch.nn.functional as F
 from flashinfer.activation import silu_and_mul
 from vllm.model_executor.custom_op import CustomOp

python/sglang/srt/layers/fused_moe/__init__.py

@@ -0,0 +1 @@
+from sglang.srt.layers.fused_moe.layer import FusedMoE, FusedMoEMethodBase