sglang 0.2.11__py3-none-any.whl → 0.2.12__py3-none-any.whl

sglang/test/test_utils.py

@@ -12,6 +12,8 @@ from typing import Callable, List, Optional
 
 import numpy as np
 import requests
+import torch
+import torch.nn.functional as F
 
 from sglang.global_config import global_config
 from sglang.lang.backend.openai import OpenAI
@@ -19,6 +21,7 @@ from sglang.lang.backend.runtime_endpoint import RuntimeEndpoint
 from sglang.utils import get_exception_traceback
 
 DEFAULT_MODEL_NAME_FOR_TEST = "meta-llama/Meta-Llama-3.1-8B-Instruct"
+DEFAULT_URL_FOR_TEST = "http://127.0.0.1:8157"
 
 
 def call_generate_lightllm(prompt, temperature, max_tokens, stop=None, url=None):
@@ -396,6 +399,8 @@ def popen_launch_server(
     timeout: float,
     api_key: Optional[str] = None,
     other_args: tuple = (),
+    env: Optional[dict] = None,
+    return_stdout_stderr: bool = False,
 ):
     _, host, port = base_url.split(":")
     host = host[2:]
@@ -415,7 +420,16 @@ def popen_launch_server(
     if api_key:
         command += ["--api-key", api_key]
 
-    process = subprocess.Popen(command, stdout=None, stderr=None)
+    if return_stdout_stderr:
+        process = subprocess.Popen(
+            command,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            env=env,
+            text=True,
+        )
+    else:
+        process = subprocess.Popen(command, stdout=None, stderr=None, env=env)
 
     start_time = time.time()
     while time.time() - start_time < timeout:
@@ -482,7 +496,7 @@ def run_unittest_files(files: List[str], timeout_per_file: float):
             p.terminate()
             time.sleep(5)
             print(
-                "\nTimeout after {timeout_per_file} seconds when running {filename}\n"
+                f"\nTimeout after {timeout_per_file} seconds when running {filename}\n"
             )
             return False
 
@@ -492,3 +506,7 @@ def run_unittest_files(files: List[str], timeout_per_file: float):
         print(f"Fail. Time elapsed: {time.time() - tic:.2f}s")
 
     return 0 if success else -1
+
+
+def get_similarities(vec1, vec2):
+    return F.cosine_similarity(torch.tensor(vec1), torch.tensor(vec2), dim=0)

sglang/utils.py

@@ -6,7 +6,6 @@ import json
 import logging
 import signal
 import sys
-import threading
 import traceback
 import urllib.request
 from concurrent.futures import ThreadPoolExecutor

sglang/version.py

@@ -1 +1 @@
-__version__ = "0.2.11"
+__version__ = "0.2.12"

{sglang-0.2.11.dist-info → sglang-0.2.12.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sglang
-Version: 0.2.11
+Version: 0.2.12
 Summary: SGLang is yet another fast serving framework for large language models and vision language models.
 License: Apache License
                                    Version 2.0, January 2004
@@ -308,7 +308,7 @@ pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
 ### Method 2: From source
 ```
 # Use the last release branch
-git clone -b v0.2.11 https://github.com/sgl-project/sglang.git
+git clone -b v0.2.12 https://github.com/sgl-project/sglang.git
 cd sglang
 
 pip install --upgrade pip
@@ -392,23 +392,23 @@ print(response)
 It supports streaming, vision, and most features of the Chat/Completions/Models/Batch endpoints specified by the [OpenAI API Reference](https://platform.openai.com/docs/api-reference/).
 
 ### Additional Server Arguments
-- Add `--tp 2` to enable tensor parallelism. If it indicates `peer access is not supported between these two devices`, add `--enable-p2p-check` option.
+- Add `--tp 2` to enable multi-GPU tensor parallelism. If it reports the error "peer access is not supported between these two devices", add `--enable-p2p-check` to the server launch command.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --tp 2
 ```
-- Add `--dp 2` to enable data parallelism. It can also be used together with tp. Data parallelism is better for throughput if there is enough memory.
+- Add `--dp 2` to enable multi-GPU data parallelism. It can also be used together with tensor parallelism. Data parallelism is better for throughput if there is enough memory.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --dp 2 --tp 2
 ```
-- If you see out-of-memory errors during serving, please try to reduce the memory usage of the KV cache pool by setting a smaller value of `--mem-fraction-static`. The default value is `0.9`.
+- If you see out-of-memory errors during serving, try to reduce the memory usage of the KV cache pool by setting a smaller value of `--mem-fraction-static`. The default value is `0.9`.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --mem-fraction-static 0.7
 ```
-- If you see out-of-memory errors during prefill for long prompts on a model that supports long context, consider using chunked prefill.
+- See [hyperparameter_tuning.md](docs/en/hyperparameter_tuning.md) on tuning hyperparameters for better performance.
+- If you see out-of-memory errors during prefill for long prompts, try to set a smaller chunked prefill size.
 ```
-python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-8B-Instruct --port 30000 --chunked-prefill-size 8192
+python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --chunked-prefill-size 4096
 ```
-- See [hyperparameter_tuning.md](docs/en/hyperparameter_tuning.md) on tuning hyperparameters for better performance.
 - Add `--nnodes 2` to run tensor parallelism on multiple nodes. If you have two nodes with two GPUs on each node and want to run TP=4, let `sgl-dev-0` be the hostname of the first node and `50000` be an available port.
 ```
 # Node 0
@@ -418,13 +418,13 @@ python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --tp 4 --nccl-init sgl-dev-0:50000 --nnodes 2 --node-rank 1
 ```
 - If the model does not have a template in the Hugging Face tokenizer, you can specify a [custom chat template](docs/en/custom_chat_template.md).
-- To enable fp8 quantization, you can add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
 - To enable experimental torch.compile support, you can add `--enable-torch-compile`. It accelerates small models on small batch sizes.
-
+- To enable fp8 quantization, you can add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
+ 
 ### Supported Models
 
 - Llama / Llama 2 / Llama 3 / Llama 3.1
-- Mistral / Mixtral
+- Mistral / Mixtral / Mistral NeMo
 - Gemma / Gemma 2
 - Qwen / Qwen 2 / Qwen 2 MoE
 - DeepSeek / DeepSeek 2
@@ -442,11 +442,20 @@ python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct
 - Grok
 - ChatGLM
 - InternLM 2
-- Mistral NeMo
 
 Instructions for supporting a new model are [here](https://github.com/sgl-project/sglang/blob/main/docs/en/model_support.md).
 
-### Run Llama 3.1 405B
+#### Use Models From ModelScope
+To use model from [ModelScope](https://www.modelscope.cn), setting environment variable SGLANG_USE_MODELSCOPE.
+```
+export SGLANG_USE_MODELSCOPE=true
+```
+Launch [Qwen2-7B-Instruct](https://www.modelscope.cn/models/qwen/qwen2-7b-instruct) Server
+```
+SGLANG_USE_MODELSCOPE=true python -m sglang.launch_server --model-path qwen/Qwen2-7B-Instruct --port 30000
+```    
+
+#### Run Llama 3.1 405B
 
 ```bash
 ## Run 405B (fp8) on a single node
@@ -474,7 +483,7 @@ GLOO_SOCKET_IFNAME=eth0 python3 -m sglang.launch_server --model-path meta-llama/
   ```
 
 ## Frontend: Structured Generation Language (SGLang)
-The frontend language can be used with local models or API models.
+The frontend language can be used with local models or API models. It is an alternative to the OpenAI API. You may found it easier to use for complex prompting workflow.
 
 ### Quick Start
 The example below shows how to use sglang to answer a mulit-turn question.

sglang-0.2.12.dist-info/RECORD

@@ -0,0 +1,112 @@
+sglang/__init__.py,sha256=T8MYdFfKFPZcgFKHMBpOCIlFbhjwmr77Nqm6mdE6bCY,1590
+sglang/api.py,sha256=gAY9JhqWXjrYoWnMvR-iiuuY1YSN94We-lc1LH0z3cw,6030
+sglang/bench_latency.py,sha256=E-cfuZSjBGonzKL0LgB0zAqMWpiP3qozB_Ht9dH8qvc,16207
+sglang/bench_serving.py,sha256=sS-fawAyzngrOVbPE3N1FBxPojoPd9vj9XQDsWpIYTQ,35798
+sglang/check_env.py,sha256=oU8VmjjPK2SviRhr41cF1953soBu-eTT5E0Hf04zMzo,4974
+sglang/global_config.py,sha256=9JxaFkBKSgep6BVeEl_kx9tuW9PqdijYELyBGTryl6o,1704
+sglang/launch_server.py,sha256=Gg8CwNlTCCfg1dF65ZT9ePLxOT9LKtY79GhIPG6PCrU,358
+sglang/launch_server_llavavid.py,sha256=40uaazMsavKuk6YXFa5v37kdUpFGuealgJJeph1g8gU,1025
+sglang/utils.py,sha256=zFYGkC4vOUR3sTv1TmQXcsOLZDtDBR3wnjqnDp3xMIs,8352
+sglang/version.py,sha256=X4KG3FscE5AhbGbcdDDgdDC550CVpxNMwdNLcx6EQ7M,23
+sglang/lang/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+sglang/lang/chat_template.py,sha256=psIlhaDo70twgLrx5Lgln03metLEA3-FZuixeI0Y7Ao,13309
+sglang/lang/choices.py,sha256=-W1DVw9N9ZliVpvmWrzIXG4cswAah8eMQrHWzkS3D8o,6234
+sglang/lang/compiler.py,sha256=1Tc6MQs4RsIfrNmmO7PMSUEHIqvNqKOp_HxaYqonwFE,7533
+sglang/lang/interpreter.py,sha256=3RIeSGdKlKTq2Ixg_Tyo0fGEDTvBKS2f9FaJYODBHzA,30102
+sglang/lang/ir.py,sha256=Ow6jXDPIeRd1piAuYjvgyFxfro1G2_-1QwUFfq4Aihs,16842
+sglang/lang/tracer.py,sha256=borJmlSJOhg1RUndGRnilnR60eEZz2Y9aU7BpftsOxU,8287
+sglang/lang/backend/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+sglang/lang/backend/anthropic.py,sha256=EXRX7xJgA5KZszX7toSLVnKzFQ5EO0Loj-YjHFtxSxg,2081
+sglang/lang/backend/base_backend.py,sha256=Q5HdiDtyBewQeoYH0kDtBRVL8KFiEPNq9dw7XmauHQ8,1985
+sglang/lang/backend/litellm.py,sha256=ugmL7sfUxkUHVbHtwNzHgdQAEd4UCjNQboFuE3KThcY,2450
+sglang/lang/backend/openai.py,sha256=qM7eVH_kMxnDd2rpxOH0v76KxtOJFlAwgLgWIKvFGCI,15060
+sglang/lang/backend/runtime_endpoint.py,sha256=AaBc5yczchX7mkwiKDMyjLjBkJsh2Lubrfd9lvCOlDo,9544
+sglang/lang/backend/vertexai.py,sha256=O-iBLD-y3vq80UxnrAoJri7bxpgd-_eakZ88Cf8bEGA,4855
+sglang/srt/conversation.py,sha256=V5YuoeO6-aLqGv0p3J2qx8TnBJbN1oTopYFutNul3GQ,16491
+sglang/srt/hf_transformers_utils.py,sha256=Tf_RplcW7llVXsigRvSGqmeAUxBeAL8rPCkzuqWfZ8U,11925
+sglang/srt/mm_utils.py,sha256=n7_GmbOM_0IWVXovpM34rKIBw0Py9yb_NXSQw27u4OA,9454
+sglang/srt/model_config.py,sha256=k4OfRV-szWkFaJMIC40JoJGJ75AfYQ2hf4M1dS1aQ-o,6366
+sglang/srt/sampling_params.py,sha256=5V1MhhEvyCWZrCF5VmQxcKNuKVoC4LynY-q4Bx3P3mo,4876
+sglang/srt/server.py,sha256=FvczPB9ojDVLIdC2kic0RLAmOTt0WZrql_BvYzwbeRY,18495
+sglang/srt/server_args.py,sha256=GLuJkgwv-Osmf3IqCvZqfdqIBJjcHkdtoNT0_zq75Kc,16849
+sglang/srt/utils.py,sha256=ReJqGMdquK_cfve269yjpWWQaozTVoEHSLG5P3CKvAg,24102
+sglang/srt/constrained/__init__.py,sha256=NLpZGj9RIx83ejDrM_pfaRtqGgaPq_ggJszPQENUJ2E,2037
+sglang/srt/constrained/base_tool_cache.py,sha256=5sazBMHHDpHMoqOjuY6itCxwTmIFCflIWEDXMtmrPVs,2006
+sglang/srt/constrained/fsm_cache.py,sha256=QTrBFoZCp2FeigtIakz2MCgQLtvQFXgl2lDPQaGtu9M,2784
+sglang/srt/constrained/jump_forward.py,sha256=IgZ8D0woy5FLIQvXkE8wZRYejDsfVkjU0sqUlkiv_f4,6193
+sglang/srt/layers/activation.py,sha256=MXkuGi5caKHEwqUegoEfOk2Omab8OLrxP-sjPj2TVzU,1197
+sglang/srt/layers/decode_attention.py,sha256=Vgxd2rWzSZkNFp0bjZRAUAusG4bz6iy3D0CULnN-cdk,8904
+sglang/srt/layers/extend_attention.py,sha256=_LOgzSr-1c2UweHZXADjWHbXOmd2JPm-tUMb1vwTTZI,14197
+sglang/srt/layers/fused_moe.py,sha256=KmyXwau2OOZpQimGIQrHptzGNs1trIud5AKEEKXdzPU,20823
+sglang/srt/layers/layernorm.py,sha256=RzN4eESN9S8mw32r2Nxarq7wKFdeG1yhxPmehUMx79s,2073
+sglang/srt/layers/logits_processor.py,sha256=iewPk7VR4jdJeLH6NAO_XqwqM4RhIHdWJzj7-qPRYIw,11362
+sglang/srt/layers/pooler.py,sha256=qNMG3Ycvt2yf9mk1Lcs-2K7oPeCuVeDYoHAxkMu9b_Q,1610
+sglang/srt/layers/prefill_attention.py,sha256=y7vdcuX8lMa9Qf_jQYNDvQO9PVCBQSs3hb5LV2DFgpU,5256
+sglang/srt/layers/radix_attention.py,sha256=LpfTizXKXm1oS5oUfh6aowZceHUHqnquvx-GpfyYjdk,7508
+sglang/srt/managers/controller_multi.py,sha256=LYI-XE9h57DW8Uh4gpd8upsC3p2dd5weKzddEH274jg,6626
+sglang/srt/managers/controller_single.py,sha256=CdQ9_XPZdcWF5jArDmVR8K-WZ9_8Gpgk4SwANKxTX-Y,5112
+sglang/srt/managers/detokenizer_manager.py,sha256=OXufjdCt2ebt-S7MDndjY9Ew16rP4fhualGgj6YEKp0,6295
+sglang/srt/managers/io_struct.py,sha256=Xvfl6DNZ2Ek2S4qlRzpVo3foc-aC-1-N-5odcJ4gdq4,9446
+sglang/srt/managers/policy_scheduler.py,sha256=KRFaZwjCAkPQDX3W8lbzrxYqgOe7LKFDj2BPlcmlnR8,8379
+sglang/srt/managers/schedule_batch.py,sha256=iZ2OwdEn5As7cVGAoe0x97cMCPSS6q_SI_iG79mF8LQ,31111
+sglang/srt/managers/tokenizer_manager.py,sha256=TIIo4YlfdM10LE4JVqv2cO2uDJJtKXDagwzfjMCDU5Q,24858
+sglang/srt/managers/tp_worker.py,sha256=qOx99QL6BIW0aOz7SknWqgflLeNeFYpJsGq0ZsYmYFY,32805
+sglang/srt/mem_cache/base_prefix_cache.py,sha256=qEQwEkG4E5rab2ZoTqcesf5pR_J4nV2jBxIHsBJHtIM,924
+sglang/srt/mem_cache/chunk_cache.py,sha256=CjZZYlqQzq7mYOiBMLWA5XNb6HIyh5lIMdY-K0OUZEc,2368
+sglang/srt/mem_cache/flush_cache.py,sha256=pTLKPRB17U6vl5RFJJvuJ4jCL2SyomgkUBNlkDpGRqo,978
+sglang/srt/mem_cache/memory_pool.py,sha256=eXDCstd5Mvu1CbHt1y9z27Eq60QYwW45FsKbZspu4yw,5310
+sglang/srt/mem_cache/radix_cache.py,sha256=0AVr1BKKDOtTyybUkwxrz6PT8khDx-DpzgN5MgL27IE,10088
+sglang/srt/model_executor/cuda_graph_runner.py,sha256=xQgTTtoMkvYJhYyRJHxPdybmPtfvcODqPLW9btUFt60,10003
+sglang/srt/model_executor/forward_batch_info.py,sha256=B3flTlRNLMa7Km7use1O0Z2YL3-a6rw1BodNKjKV51g,11049
+sglang/srt/model_executor/model_runner.py,sha256=ZlFgqBNuqgWpa-NrjkfTT-_amtea33H9M1tBl-MT_nk,16977
+sglang/srt/model_loader/model_loader.py,sha256=QmZUhHh1nmWrfYlunfnxMcTsIvip1l6aMIlrXoCED4I,10697
+sglang/srt/model_loader/utils.py,sha256=0AoWXX9uV5rKRYXJ4HduSnvdeerytI4ONCLCH6X4XFQ,10675
+sglang/srt/models/chatglm.py,sha256=7bHU2AFoppINDZm0EdxgtAJe7rwr9OPkhOCfq2qNrIA,13862
+sglang/srt/models/commandr.py,sha256=5BEtIS2uUQJANkkY-6ZeDqlrpUK5yXVYHiztU3vsTKY,14172
+sglang/srt/models/dbrx.py,sha256=N_0Ku_p1NCsc29NktUBNqPv7Z33XhYxOZK5xN7nzW4s,14661
+sglang/srt/models/deepseek.py,sha256=E5W4nkH-Ne449rAIwQZgz-FAH2Qqp2r1vNfboyk5wEg,16024
+sglang/srt/models/deepseek_v2.py,sha256=NMcckZb48kVUwAmDA2l8wO19T6DNkJOkKAhHa6utBZM,26968
+sglang/srt/models/gemma.py,sha256=ilfN_NOcz7hpwEJ2y7NW3fBFmFO7YfjhdFDbfzl2qww,12285
+sglang/srt/models/gemma2.py,sha256=ybQOXAPofw_Pv3mBer7dTpH4SlZt6Gf2I462Q3lOIww,16359
+sglang/srt/models/gpt_bigcode.py,sha256=OKk9UP67as3T5bePlTRGHTCD-1wqaUEk92AowXPm6dg,10204
+sglang/srt/models/grok.py,sha256=M9rtdXslqYBle5VyZqFVHiJUXq_q_aHbza63xa03zqI,27861
+sglang/srt/models/internlm2.py,sha256=6j7JH0p3yib8GZDH8Cmrs-pgwfH3eOlAK6V3Cq64O7w,12202
+sglang/srt/models/llama2.py,sha256=HmzE1I8OnesmrdPY5b56l7okhWH_lRvWAg16K-UwKHg,14300
+sglang/srt/models/llama_classification.py,sha256=Dvzy3PfETiJtnKFOk8qDDLUoZECf_cpSrNeA60PaDo4,4932
+sglang/srt/models/llama_embedding.py,sha256=e2lpZ6GHKrHT1rr7_5gHGoCpfqdOBMusZCz34n62lec,3542
+sglang/srt/models/llava.py,sha256=-ysi192vpBDxNaMS8qaLOhC34lXQyRtbG_0niVaceSo,18436
+sglang/srt/models/llavavid.py,sha256=MX7YpqYh5J4BoOnV7vVAIfoOlBFQXYpp8Kpe7WK0ejk,13562
+sglang/srt/models/minicpm.py,sha256=ea_OyiwVTo6Tg9jNRAwqxETnA6FFeAqlIbiUS-xViEI,13843
+sglang/srt/models/mistral.py,sha256=jlrWBVNXbAUziAaIdHAjFcOJnKtn9Bl8rBd65ypJM-I,819
+sglang/srt/models/mixtral.py,sha256=raSLbp6AfWg5_u-f-lYeRejE9koAjbHt8iIHXd3nURM,21397
+sglang/srt/models/mixtral_quant.py,sha256=xYeeatZ9OfwCTas_KbH9nl6lnUT4YqSY7NAxpgLp5LE,14222
+sglang/srt/models/qwen.py,sha256=43ea6gn4wHzAaI3JTDLtl08aEm0vIqgzbVH9M8oeuY0,10006
+sglang/srt/models/qwen2.py,sha256=Hyhks2r4KHpKeb9iHZpnvEVc5klmnrPwcLohqg8j1kw,12284
+sglang/srt/models/qwen2_moe.py,sha256=pTfBivDyzdbcP22_7PdmdPqgx34esH8J98r-EgFA9Uw,17747
+sglang/srt/models/stablelm.py,sha256=yPrdzPEoUD2s_Q3RgOq7BBC7z-UtEaACzabqbDRs2tA,11368
+sglang/srt/models/yivl.py,sha256=p4s_D_m4H2exP4b91Y-CTkq8T-eIG3DJsFy9pB0e7TM,4932
+sglang/srt/openai_api/adapter.py,sha256=fgUAPAcQ_mUJszbpsI_cgv2vzOAS7AKKAJPi2B91aw4,42490
+sglang/srt/openai_api/protocol.py,sha256=knf-nds0XO2LYg-hPM-Ho1f1y2XZIV_Gvg3xcCKLfgQ,9411
+sglang/srt/sampling/penaltylib/__init__.py,sha256=5vQw0Y5DSzmsoFg1IdMIKLwFVhYZ5ArADHVBYbSmOec,513
+sglang/srt/sampling/penaltylib/orchestrator.py,sha256=WkTNeDhj9H9rtp2ZZeX6MS2sdKSGlLboE6FcuKrwUo0,10815
+sglang/srt/sampling/penaltylib/penalizers/frequency_penalty.py,sha256=IvYioX53Vq_ji-0Zhcz_r5mUa3T3GaIydVS6K4FhWfE,2557
+sglang/srt/sampling/penaltylib/penalizers/min_new_tokens.py,sha256=XJZP0C4NFyXgcODbIWXxrgVEjmRgqLdZuVAtoN-LveY,3565
+sglang/srt/sampling/penaltylib/penalizers/presence_penalty.py,sha256=0PlANTrR959foTA3Nj5qBE7ndaOZgG-9X6LhzlmEUc8,2533
+sglang/srt/sampling/penaltylib/penalizers/repetition_penalty.py,sha256=v9jOgA0-I31WcrhIydiFbpy2ZJPLytFLGM98NRPd2sU,2820
+sglang/test/run_eval.py,sha256=NWxeLWmInBgkCvC9Jr_QzF7GfAiBve3Gf1JQrEOlNlU,3899
+sglang/test/runners.py,sha256=FYLbrWePfTacN5bsbAgMl5RiDI4g_Bsbwh1gXqRwr0Y,7794
+sglang/test/simple_eval_common.py,sha256=HL1bfgkTAKP7sk-kShg73WTeADhuBD6xSsuLbV_9C3s,12359
+sglang/test/simple_eval_gpqa.py,sha256=CaRAuHdZj0m4mRm4tH9k7cB0kQxe0LHwlz7Vn1qyKps,3189
+sglang/test/simple_eval_humaneval.py,sha256=iCtN2LBL6j3nxMDjRJ--m0MCNPAwDo81gJ2whE-2Rt0,5674
+sglang/test/simple_eval_math.py,sha256=EQblQmtUt-kl558drzhP7c6KhpDNgr1EJhhKx5eeHM4,2519
+sglang/test/simple_eval_mgsm.py,sha256=wfbqJW9Rkc66vzq2fEMF6jchmoA8mw1OUiGU55cZ2B0,10261
+sglang/test/simple_eval_mmlu.py,sha256=KqSSdSu2qfoKQ870ttxev1NJ7c90xv2mvKOQsSODtAw,4326
+sglang/test/test_layernorm.py,sha256=VDdoeqGvebUa-l3rDiid6cC7wZq0Phpbm5fxxD0-cpg,1910
+sglang/test/test_programs.py,sha256=vRhKIriZgSk_Zn8gGviIfiY_suOBA7Ni7P0NaQM2Esk,13894
+sglang/test/test_utils.py,sha256=cO0ZbnfBS_MxyZ6MDyA7DrDVwu3umKRb3WP_dwggPng,14505
+sglang/test/srt/sampling/penaltylib/utils.py,sha256=-0p0rV-P4lNo7xAe3rQSBHTubc50a-DFyOQmLGAkgkQ,12515
+sglang-0.2.12.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+sglang-0.2.12.dist-info/METADATA,sha256=k4QBFP1vyWHeXgCA9Npoz7Wb8qT9aC8rL7R1QP2J60g,34314
+sglang-0.2.12.dist-info/WHEEL,sha256=R0nc6qTxuoLk7ShA2_Y-UWkN8ZdfDBG2B6Eqpz2WXbs,91
+sglang-0.2.12.dist-info/top_level.txt,sha256=yxhh3pYQkcnA7v3Bg889C2jZhvtJdEincysO7PEB09M,7
+sglang-0.2.12.dist-info/RECORD,,