HappyZ happyz
0 Followers · 0 Following
Repositories
18
 Projects Packages Public Activity Starred Repositories
happyz synced commits to refs/pull/896/merge at happyz/gemma.cpp from mirror 2026-04-16 09:19:58 -07:00
fb0e5fa002 Merge c4d14dbc22 into 221d8df516
c4d14dbc22 Use HWY_ABORT instead of std::cerr for overflow checks in ReadPPM
e260706346 Fix integer overflow in tensor dimensions and PPM parser
221d8df516 Increased max_tbatch_size to kMaxBatchSize. Gives 1.5x speed-up for prefill on both intel and AMD machines
a29e2fc655 Fixed bug in PackedBytes - not using override_rows.
Compare 139 commits »
happyz synced commits to refs/pull/896/head at happyz/gemma.cpp from mirror 2026-04-16 09:19:57 -07:00
c4d14dbc22 Use HWY_ABORT instead of std::cerr for overflow checks in ReadPPM
e260706346 Fix integer overflow in tensor dimensions and PPM parser
221d8df516 Increased max_tbatch_size to kMaxBatchSize. Gives 1.5x speed-up for prefill on both intel and AMD machines
a29e2fc655 Fixed bug in PackedBytes - not using override_rows.
366b143fbf Merge pull request #891 from texasich:clean-gcc15-fix
Compare 138 commits »
happyz synced new reference test_900671724 to happyz/gemma.cpp from mirror 2026-04-16 09:19:57 -07:00
happyz synced commits to test_900671724 at happyz/gemma.cpp from mirror 2026-04-16 09:19:56 -07:00
happyz synced new reference refs/tags/b8815 to happyz/llama.cpp from mirror 2026-04-16 09:19:44 -07:00
happyz synced commits to refs/tags/b8815 at happyz/llama.cpp from mirror 2026-04-16 09:19:44 -07:00
happyz synced new reference refs/tags/b8814 to happyz/llama.cpp from mirror 2026-04-16 09:19:43 -07:00
happyz synced commits to refs/tags/b8814 at happyz/llama.cpp from mirror 2026-04-16 09:19:43 -07:00
happyz synced commits to refs/tags/b8813 at happyz/llama.cpp from mirror 2026-04-16 09:19:43 -07:00
happyz synced new reference refs/tags/b8813 to happyz/llama.cpp from mirror 2026-04-16 09:19:43 -07:00
happyz synced commits to refs/tags/b8809 at happyz/llama.cpp from mirror 2026-04-16 09:19:42 -07:00
happyz synced new reference refs/tags/b8809 to happyz/llama.cpp from mirror 2026-04-16 09:19:42 -07:00
happyz synced commits to refs/tags/b8811 at happyz/llama.cpp from mirror 2026-04-16 09:19:42 -07:00
happyz synced new reference refs/tags/b8811 to happyz/llama.cpp from mirror 2026-04-16 09:19:42 -07:00
happyz synced commits to refs/pull/21977/merge at happyz/llama.cpp from mirror 2026-04-16 09:19:41 -07:00
a791262a67 Merge c5ca739ea2 into f772f6e434
f772f6e434 model : support NVFP4 tensors for Gemma4 (#21971)
b572d1ecd6 codeowners: add team member comments (#21714)
03b3d07798 Convert: Fix NemotronH Config Parsing (#21664)
3f7c29d318 ggml: add graph_reused (#21764)
Compare 12 commits »
happyz synced commits to refs/pull/21972/merge at happyz/llama.cpp from mirror 2026-04-16 09:19:40 -07:00
8abb04d6ca Merge 7c346be6bf into b572d1ecd6
b572d1ecd6 codeowners: add team member comments (#21714)
03b3d07798 Convert: Fix NemotronH Config Parsing (#21664)
3f7c29d318 ggml: add graph_reused (#21764)
ae2d34899e metal: Implement ROLL op (#21946)
Compare 11 commits »
happyz synced commits to refs/pull/21974/merge at happyz/llama.cpp from mirror 2026-04-16 09:19:40 -07:00
d42d67ccf9 Merge d70d466266 into f772f6e434
f772f6e434 model : support NVFP4 tensors for Gemma4 (#21971)
b572d1ecd6 codeowners: add team member comments (#21714)
03b3d07798 Convert: Fix NemotronH Config Parsing (#21664)
3f7c29d318 ggml: add graph_reused (#21764)
Compare 12 commits »
happyz synced commits to refs/pull/21969/merge at happyz/llama.cpp from mirror 2026-04-16 09:19:39 -07:00
f4ba5b5339 Merge 65b8574170 into 03b3d07798
03b3d07798 Convert: Fix NemotronH Config Parsing (#21664)
3f7c29d318 ggml: add graph_reused (#21764)
ae2d34899e metal: Implement ROLL op (#21946)
1e796eb41f ggml-cpu: add 128-bit RVV implementation for Quantization Vector Dot (#20633)
Compare 10 commits »
happyz synced commits to refs/pull/21970/merge at happyz/llama.cpp from mirror 2026-04-16 09:19:39 -07:00
868364130c Merge be46a5096f into b572d1ecd6
b572d1ecd6 codeowners: add team member comments (#21714)
03b3d07798 Convert: Fix NemotronH Config Parsing (#21664)
3f7c29d318 ggml: add graph_reused (#21764)
ae2d34899e metal: Implement ROLL op (#21946)
Compare 11 commits »
happyz synced commits to refs/pull/21964/merge at happyz/llama.cpp from mirror 2026-04-16 09:19:38 -07:00
a984d87457 Merge 17c99108e6 into 9db77a020c
17c99108e6 merge: merge with upstream master
9db77a020c model : refactor QKV into common build_qkv and create_tensor_qkv helpers (#21245)
f772f6e434 model : support NVFP4 tensors for Gemma4 (#21971)
b572d1ecd6 codeowners: add team member comments (#21714)
Compare 15 commits »