Ecosyste.ms: Timeline
Browse the timeline of events for every public repo on GitHub. Data updated hourly from GH Archive.
sixsixcoder created a comment on an issue on THUDM/GLM-4-Voice
不支持，GLM-4-Voice是基于GLM-4训练而来，其输入和输出都做了严格的格式化
sixsixcoder created a comment on an issue on THUDM/GLM-4
> 您好，ollama run glm4 下载的是那个模型呢？怎么指定下载glm-4-9b-chat这个版本呢 您可以事先下载`glm-4-9b-chat`到本地，或者ollama应该有地方可以设置下载模型的id
sixsixcoder created a comment on an issue on THUDM/GLM-4
更多openai参数的用法可以google搜索openai sdk的文档，那里会有详细介绍https://platform.openai.com/
sixsixcoder created a comment on an issue on THUDM/GLM-4
在`openai_api_request.py`修改`max_tokens`即可 ``` response = client.chat.completions.create( model="glm-4", messages=messages, stream=use_stream, max_tokens=256, ...
sixsixcoder created a comment on an issue on THUDM/GLM-4-Voice
感谢您的关注，这个具体没有统计，常见的方言您可以尝试一下，之前好像有人少数民族语言也支持
sixsixcoder created a comment on an issue on THUDM/GLM-4-Voice
感谢您的关注，您可以尝试一下，比如设计几个音频示例，分别为开心、生气等情绪的音频，输入到模型demo中识别
sixsixcoder created a comment on an issue on THUDM/GLM-4-Voice
[GLM-4-Voice](https://github.com/THUDM/GLM-4-Voice)是开源基础语音模型，支持语音文件输入和语音录制输入，并没有进行产品包装，只是提供一个基础的开源demo，更多的使用场景可以由开发者自由定制和二次开发。
sixsixcoder created a comment on an issue on THUDM/GLM-4
感谢您的关注，但是这个问题感觉偏理论，我觉得你可以直接联系论文作者，发送到邮箱讨论该问题
sixsixcoder created a comment on an issue on THUDM/GLM-4
Although this PR can work, it is not suitable for the GLM-4v model. We are rewriting and adding chat_template
sixsixcoder created a comment on an issue on THUDM/GLM-4
openai_api_server.py 不支持lora权重加载，你可以尝试使用finetune_demo/inference.py 进行推理
sixsixcoder created a comment on an issue on THUDM/GLM-4
流式输出可以参考这个代码：https://github.com/THUDM/GLM-4/blob/main/basic_demo/trans_cli_demo.py
sixsixcoder pushed 40 commits to glm-4 sixsixcoder/vllm
- [Model] Support `is_causal` HF config field for Qwen2 model (#10621) Signed-off-by: DarkLight1337 <tlleungac@connect... ed46f14
- Super tiny little typo fix (#10633) 2b0879b
- [Bug]: Authorization ignored when root_path is set (#10606) Signed-off-by: chaunceyjiang <[email protected]> d04b13a
- [Bugfix] Fix chunked prefill with model dtype float32 on Turing Devices (#9850) Signed-off-by: Wallas Santos <wallas... c27df94
- [Docs] Add Snowflake Slides (#10641) Signed-off-by: simon-mo <[email protected]> 452a4e8
- [Model]: Add support for Aria model (#10514) Signed-off-by: xffxff <[email protected]> Co-authored-by: Isotr0py <203... b1d9205
- [Model] Enable optional prefix when loading embedding models (#10639) Signed-off-by: DarkLight1337 <tlleungac@connec... cf73f0c
- [Doc] Fix typos in docs (#10636) Signed-off-by: DarkLight1337 <[email protected]> 1b583cf
- [Model] Add OLMo November 2024 model (#10503) 9db713a
- [misc] do not read HOST_IP (#10644) Signed-off-by: youkaichao <[email protected]> 6e9ff05
- [bugfix] fix aria model and add torch.compile (#10645) Signed-off-by: youkaichao <[email protected]> 45ac4ff
- [Feature] vLLM ARM Enablement for AARCH64 CPUs (#9228) Signed-off-by: Sanket Kale <[email protected]> Co-aut... a6760f6
- [v1] EngineArgs for better config handling for v1 (#10382) Signed-off-by: rickyx <[email protected]> 519e8e4
- custom allreduce + torch.compile (#10121) Signed-off-by: youkaichao <[email protected]> Co-authored-by: youkaich... 9a88f89
- [Misc] Remove outdated init protocols (#10655) Signed-off-by: DarkLight1337 <[email protected]> 9406353
- [ci] add vllm_test_utils (#10659) Signed-off-by: youkaichao <[email protected]> 334d64d
- [V1] Enable profile for LLMEngine (#10665) 1f6584e
- [Bugfix] Fix for Spec model TP + Chunked Prefill (#10232) Signed-off-by: andoorve <[email protected]... db66e01
- [Hardware][NVIDIA] Add non-NVML CUDA mode for Jetson (#9735) Signed-off-by: Conroy Cheers <[email protected]> f5792c7
- [Bugfix] Fix using `-O[0,3]` with LLM entrypoint (#10677) Signed-off-by: mgoin <[email protected]> 9a99273
- and 20 more ...
sixsixcoder pushed 1 commit to glm-4 sixsixcoder/vllm
- Update vllm/model_executor/models/glm.py Co-authored-by: Cyrus Leung <[email protected]> cb127dc
sixsixcoder created a comment on a pull request on vllm-project/vllm
> Hey @sixsixcoder, I just made a refactor on your GLM implementation to directly inherit from `LlamaForCausalLM` just like Phi-3, so that we can significantly reduce duplicated codes and support f...
sixsixcoder pushed 306 commits to main sixsixcoder/sglang
- Fix prefill oom (#1743) efb099c
- Faster overlap mode scheduler (#1738) 7ce3606
- misc: add CODEOWNERS (#1737) e68b9e7
- Fix sliding window attention and gemma-2 unit tests in CI (#1746) 0061128
- Llama3.2 vision model support (#1551) 94cde10
- Update `max_req_len` and `max_req_input_len` (#1748) 5e1558f
- Release v0.3.4.post1 (#1749) 1f26e8b
- Fix edge case for truncated (#1747) 17536e7
- Fuse more ops & Simplify token mapping (#1758) ad4125d
- [API] add get memory pool size (#1760) Co-authored-by: Byron Hsu <[email protected]> 2fce449
- Fix perf regression for set_kv_buffer (#1765) fbcbb26
- [Fix] Fix abort in dp (#1767) 9af7b88
- Fix stop condition for <|eom_id|> (#1766) 80a9054
- Update docs (#1768) Co-authored-by: Chayenne Zhao <[email protected]> Co-authored-by: Chayenne <zhaochen20@o... b7d0559
- Fix missing additional_stop_token_ids (#1769) 0d80009
- Fix out of memory message. (#1771) 3f5ac88
- Crash the server on warnings in CI (#1772) 05b3bf5
- Fix the perf regression due to additional_stop_token_ids (#1773) 8f8f96a
- Fix MockTokenizer in the unit tests (#1774) 87a7cfa
- [Bug] Catch any errors caused by parsing json schema (#1776) 72e7b57
- and 286 more ...
sixsixcoder created a review comment on a pull request on vllm-project/vllm
Modifications have been submitted according to your commit.