Pulse Contributors Code Frequency Recent Commits

2026-09-23 - 2026-10-23

Period: 1 month

1 day 3 days 1 week 1 month 3 months 6 months 1 year

Overview

0 Active Pull Requests

18 Active Issues

0
Merged Pull Requests 0
Proposed Pull Requests 18
Closed Issues 15
New Issues

Excluding merges, 1 author has pushed 20 commits to main and 20 commits to all branches. On main, 408 files have changed and there have been 74967 additions and 2033 deletions.

18 Issues closed from 1 user

Closed #100 support deepseek v4 flash vision experimental 2026-10-04 18:33:20 +00:00

Closed #117 Match Qwen direct-mode reasoning history rendering 2026-10-04 15:38:49 +00:00

Closed #116 Apply the Qwen tokenizer NFC normalization contract 2026-10-04 15:14:22 +00:00

Closed #115 Quote numeric-looking revisions in the Dev Brain Git fixture 2026-10-04 15:01:06 +00:00

Closed #114 Preserve Qwen prefill cached-key zero sign 2026-10-04 14:56:06 +00:00

Closed #113 Preserve MTPLX head-chunked Qwen verify SDPA execution 2026-10-04 14:45:51 +00:00

Closed #112 Port MTPLX row-map SwiGLU expert gather for Qwen prefill MoE 2026-10-04 14:25:29 +00:00

Closed #111 Tokenize all Unicode number categories with the Qwen3 contract 2026-10-04 13:37:37 +00:00

Closed #110 Prevent uninitialized expert outputs for large unaligned M5 prefill chunks 2026-10-04 13:18:52 +00:00

Closed #109 Preserve compiled BF16 attention-gate math in MTP verification 2026-10-04 12:50:58 +00:00

Closed #108 Apply M-RoPE row by row across the vision position-table boundary 2026-10-04 12:11:47 +00:00

Closed #107 Preserve sparse QSA and image-aware indexer positions for vision turns 2026-10-04 12:06:01 +00:00

Closed #106 Commit the final MTP primary before banking warm session state 2026-10-04 11:42:37 +00:00

Closed #105 Reject invalid Qwen logits instead of emitting fabricated tokens 2026-10-04 11:29:45 +00:00

Closed #104 Use full-vocabulary top-p law in Qwen AR pipeline 2026-10-04 10:50:18 +00:00

Closed #102 opening GLM 5.3 flash session when qwen is selected gives an error 2026-10-04 10:27:11 +00:00

Closed #103 Analyse Qwen3.8 Flash Next changes in MTPLX and create confirmed bug issues 2026-10-04 06:28:57 +00:00

Closed #101 Support switching model via direct dropdown on model name in chat while still empty 2026-10-02 08:35:38 +00:00

15 Issues created by 0 users

Opened #103 Analyse Qwen3.8 Flash Next changes in MTPLX and create confirmed bug issues 2026-10-02 20:52:13 +00:00

Opened #104 Use full-vocabulary top-p law in Qwen AR pipeline 2026-10-03 09:20:30 +00:00

Opened #105 Reject invalid Qwen logits instead of emitting fabricated tokens 2026-10-03 09:20:31 +00:00

Opened #106 Commit the final MTP primary before banking warm session state 2026-10-03 09:20:31 +00:00

Opened #107 Preserve sparse QSA and image-aware indexer positions for vision turns 2026-10-03 09:20:32 +00:00

Opened #108 Apply M-RoPE row by row across the vision position-table boundary 2026-10-03 09:20:32 +00:00

Opened #109 Preserve compiled BF16 attention-gate math in MTP verification 2026-10-03 09:20:32 +00:00

Opened #110 Prevent uninitialized expert outputs for large unaligned M5 prefill chunks 2026-10-03 09:20:33 +00:00

Opened #111 Tokenize all Unicode number categories with the Qwen3 contract 2026-10-03 09:20:33 +00:00

Opened #112 Port MTPLX row-map SwiGLU expert gather for Qwen prefill MoE 2026-10-03 17:09:37 +00:00

Opened #113 Preserve MTPLX head-chunked Qwen verify SDPA execution 2026-10-04 12:33:12 +00:00

Opened #114 Preserve Qwen prefill cached-key zero sign 2026-10-04 12:46:37 +00:00

Opened #115 Quote numeric-looking revisions in the Dev Brain Git fixture 2026-10-04 13:17:46 +00:00

Opened #116 Apply the Qwen tokenizer NFC normalization contract 2026-10-04 13:33:42 +00:00

Opened #117 Match Qwen direct-mode reasoning history rendering 2026-10-04 13:33:53 +00:00