Show Harden Runner enabled runs only
Merge pull request #980 from Mikyx-1:cmake-fetch-progress-dev
dev
build #7842build.yml3 days ago
5Destinations
7Actions
Insights
Merge pull request #981 from RaviTriv:fix-softmax-temperature
dev
build #7841build.yml3 days ago
5Destinations
7Actions
Insights
Deduplicate packed matrix splitting
dev
build #7825build.yml5 days ago
5Destinations
4Actions
Insights
Internal Changes
dev
build #7812build.yml8 days ago
5Destinations
4Actions
Insights
Gemma 4 E2B Vision
dev
build #7786build.yml10 days ago
5Destinations
4Actions
Insights
Merge pull request #970 from Intel-tensorflow:feature/onednn-matmul-threadpool
dev
build #7780build.yml10 days ago
5Destinations
7Actions
Insights
Dspark support for converter and inference
dev
build #7772build.yml11 days ago
5Destinations
4Actions
Insights
Fix buffer overflow in compression int_test.
dev
build #7763build.yml13 days ago
5Destinations
4Actions
Insights
Serialization: EnumValid uses switch, not sentinel; add framing
dev
build #7746build.yml17 days ago
5Destinations
7Actions
Insights
Optimize Gemma 4 MoE router input buffering and scaling:
dev
build #7697build.yml21 days ago
5Destinations
4Actions
Insights
Gemma4 can now use tiled attention implementation
dev
build #7621build.yml28 days ago
5Destinations
4Actions
Insights
[gemma_cpp] Add support for Qwen3 models (0.6B, 1.7B and 4B).
dev
build #7616build.yml28 days ago
5Destinations
4Actions
Insights
Internal change.
dev
build #7596build.yml28 days ago
5Destinations
4Actions
Insights
Merge pull request #947 from maasir554:t5gemma-support
dev
build #7568build.ymlabout 1 month ago
5Destinations
7Actions
Insights
Merge pull request #963 from RaviTriv:profiler-build-flag
dev
build #7567build.ymlabout 1 month ago
5Destinations
7Actions
Insights
Implement VNNI based flash attention that uses SumOfMulQuadAccumulate
dev
build #7551build.ymlabout 1 month ago
5Destinations
4Actions
Insights
Pass runtime_config_ to the KVCache constructor within GemmaEnv in benchmark_helper.cc. This ensures that kv_cache_type and attention_impl settings are properly propagated when using non-default attention implementations such as flash_transposed_qs_bf16.
dev
build #7495build.ymlabout 1 month ago
5Destinations
4Actions
Insights
Gemma 4 E2B text model
dev
build #7477build.ymlabout 1 month ago
5Destinations
4Actions
Insights
Add CONTRIBUTING
dev
build #7461build.ymlabout 1 month ago
5Destinations
4Actions
Insights
Adjust work spliting to subtasks, about 1.5x faster prefill on zen5 machines, and +- neutral on older ones. Importantly to fully utilize this it's important to use high prefill token_batch like 4k
dev
build #7454build.ymlabout 1 month ago
5Destinations
5Actions
Insights
For flash_tiled_attention... fixed support for models with mixed qkv dim.
dev
build #7449build.ymlabout 1 month ago
5Destinations
4Actions
Insights
Better spreading of work for tiled attention
dev
build #7403build.ymlabout 2 months ago
5Destinations
4Actions
Insights
Add kv data type flag
dev
build #7394build.ymlabout 2 months ago
5Destinations
4Actions
Insights
Automated Code Change
dev
build #7384build.ymlabout 2 months ago
5Destinations
4Actions
Insights
Move non-SIMD TwoSum into fp_arith.h for use by fp_arith-inl.h
dev
build #7363build.ymlabout 2 months ago
5Destinations
4Actions
Insights
Also namespace-prefix in dot_test.cc
dev
build #7339build.ymlabout 2 months ago
5Destinations
4Actions
Insights
Namespace-qualify all calls to fp_arith-inl.h so we can move it to hwy
dev
build #7328build.ymlabout 2 months ago
5Destinations
4Actions
Insights
Arm bfmmla based flash attention kernel
dev
build #7313build.ymlabout 2 months ago
5Destinations
4Actions
Insights
Merge pull request #940 from qiuku30:fix/kvcache-topt-missing-cache-ptr
dev
build #7287build.yml2 months ago
5Destinations
7Actions
Insights
[MatMul] Increase reps while autotuning for SFC traversals in MatMul.
dev
build #7269build.yml2 months ago
5Destinations
4Actions
Insights