Skip to content

Pull requests: google/tunix

New pull request New
20 Open 101 Closed
20 Open 101 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

move sync_weight to rl_cluster
#126 opened Jul 21, 2025 by copybara-service bot Loading…
Fix Qwen3 attention
#125 opened Jul 21, 2025 by copybara-service bot Loading…
1
remove unnecessary comments.
#124 opened Jul 21, 2025 by copybara-service bot Loading…
improve test coverage.
#123 opened Jul 21, 2025 by copybara-service bot Loading…
[Tunix] Add support for hooks in training loop.
#119 opened Jul 18, 2025 by copybara-service bot Loading…
[TunixW&B] Install wandb in qlora_demo
#113 opened Jul 17, 2025 by copybara-service bot Loading…
Change default sharding to (2, 4) for GRPO demo.
#111 opened Jul 16, 2025 by copybara-service bot Loading…
enable resharding cm for only tpu devices
#108 opened Jul 15, 2025 by copybara-service bot Loading…
Improve Llama3 loading
#106 opened Jul 14, 2025 by copybara-service bot Loading…
1
Fix pad id and add a unit test for the tokenizer adapter.
#105 opened Jul 14, 2025 by copybara-service bot Loading…
1
Add the get_model_input method to Gemma3 for returning a dummy input CLA:yes
#101 opened Jul 11, 2025 by yhtang Loading…
grpo pathway demo fix.
#95 opened Jul 7, 2025 by copybara-service bot Loading…
1
Add Qwen3 from_pretrained, generation example, and DeepseekV3 model support
#92 opened Jul 3, 2025 by danikhan632 Loading…
Preparation for fp8 quantized training support
#91 opened Jul 3, 2025 by copybara-service bot Loading…
measure reshard time
#85 opened Jun 30, 2025 by copybara-service bot Loading…
Adding PPO trainer
#80 opened Jun 27, 2025 by danikhan632 • Draft
1
Change nnx split and store transformer state.
#74 opened Jun 18, 2025 by copybara-service bot Loading…
1
Load model on the workers CPU to mitigate the resource stress of the client.
#73 opened Jun 12, 2025 by copybara-service bot Loading…
1
Add an optional section to serve the GRPO-tuned model via vLLM in the GRPO tutorial
#18 opened May 13, 2025 by copybara-service bot Loading…
Add files via upload
#16 opened May 10, 2025 by charbull Loading…
ProTip! Filter pull requests by the default branch with base:main.