Wallas Henrique
Wallas Henrique
When changes the orientation of the device to portrait, the screen freezes. Sounds and controls still work. Tested on a iPhone 5.1 and Xcode 4.3.
### Your current environment The output of `python collect_env.py` ```text Collecting environment information... INFO 09-13 19:13:45 importing.py:10] Triton not installed; certain GPU-related functions will not be available. PyTorch version: 2.4.0+cpu...
### Your current environment The output of `python collect_env.py` ```text Collecting environment information... INFO 09-13 19:13:45 importing.py:10] Triton not installed; certain GPU-related functions will not be available. PyTorch version: 2.4.0+cpu...
### Your current environment The output of `python collect_env.py` ```text $ python collect_env.py Collecting environment information... /opt/conda/envs/py_3.9/lib/python3.9/site-packages/transformers/utils/hub.py:127: FutureWarning: Using `TRANSFORMERS_CACHE` is deprecated and will be removed in v5 of Transformers....
## Don't log duplicate error stacktrace for every request in the batch **EDIT:** Discussed with @joerunde, and we changed the solution. On the `client.py` if the engine is errored, then...
Greetings, We did a study of mutual exclusive features on vLLM and consolidated in a compatibility matrix. We propose to add the compatibility matrix to the documentation pages to help...
[Bug]: Speculative decoding generate gibberish when receiving parallel requests with different seeds
### Your current environment The output of `python collect_env.py` ```text Collecting environment information... PyTorch version: 2.4.0+cu121 Is debug build: False CUDA used to build PyTorch: 12.1 ROCM used to build...
Mistral models + guided decoding with json schema with xgrammar is generating endless whitespace. This bug was introduced with this [change](https://github.com/mlc-ai/xgrammar/commit/7cc2814389925c7a3d98730304e63aa39169cf1d#diff-14b813481d67a86d31641fb13b641bffe32924eca2054e61b30fd2c6299d96d6R82) on xgrammar. My proposal is to add an environment...
This PR adds a feature to dump input metadata when vllm engine crashes. In essence, this change is the spiritual successor to https://github.com/vllm-project/vllm/pull/8305 that was recently removed in https://github.com/vllm-project/vllm/pull/12582. However,...
This small patch allows to users have an "empty install" on mac os if they wish so. Currently, any install on mac os force to CPU.