Skip to content

Actions: ml-explore/mlx-swift-lm

Actions

Build and Test

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
pull_request.yml
486 workflow runs
486 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Add Gemma 4 12B unified Build and Test #784: Pull request #327 synchronize by adrgrondin
Action required adrgrondin:model/gemma-4-12b-unified
adrgrondin:model/gemma-4-12b-unified
Action required
fix MLXVLM prepare(): honor windowSize for chunked prefill (FastVLM / Gemma3 / Qwen2VL / LFM2VL / Pixtral / Mistral3) Build and Test #783: Pull request #344 synchronize by davidkoski
26m 11s john-rocky:fix/vlm-prepare-window-size
john-rocky:fix/vlm-prepare-window-size
26m 11s
Add parser for GPT-OSS Harmony tool call format Build and Test #782: Pull request #146 synchronize by aleroot
Action required aleroot:harmony_tools
aleroot:harmony_tools
Action required
Add variance-normalized KV cache Build and Test #781: Pull request #329 synchronize by aleroot
Action required aleroot:kvarn-inspired-kv-cache
aleroot:kvarn-inspired-kv-cache
Action required
Base Gemma diffusion implementation Build and Test #780: Pull request #352 synchronize by aleroot
12m 2s aleroot:gemma_diffusion
aleroot:gemma_diffusion
12m 2s
Add Qwen3.5 MTP speculative decoding Build and Test #779: Pull request #351 synchronize by aleroot
Action required aleroot:Qwen3MTP
aleroot:Qwen3MTP
Action required
Improve Qwen3.5 recurrent cache handling Build and Test #778: Pull request #323 synchronize by aleroot
5m 44s aleroot:qwen3_5perf
aleroot:qwen3_5perf
5m 44s
Improve Qwen3.5 recurrent cache handling Build and Test #777: Pull request #323 synchronize by aleroot
8m 35s aleroot:qwen3_5perf
aleroot:qwen3_5perf
8m 35s
Improve Qwen3.5 recurrent cache handling Build and Test #776: Pull request #323 synchronize by aleroot
13m 54s aleroot:qwen3_5perf
aleroot:qwen3_5perf
13m 54s
Add Gemma 4 12B unified Build and Test #775: Pull request #327 synchronize by adrgrondin
16m 38s adrgrondin:model/gemma-4-12b-unified
adrgrondin:model/gemma-4-12b-unified
16m 38s
Fix Gemma4 E-series load: make per_layer_model_projection quantizable Build and Test #774: Pull request #320 synchronize by davidkoski
13m 56s nyteshade:fix/gemma4-ple-quantization
nyteshade:fix/gemma4-ple-quantization
13m 56s
Base Gemma diffusion implementation Build and Test #773: Pull request #352 opened by aleroot
Action required aleroot:gemma_diffusion
aleroot:gemma_diffusion
Action required
Add Qwen3.5 MTP speculative decoding Build and Test #772: Pull request #351 synchronize by aleroot
Action required aleroot:Qwen3MTP
aleroot:Qwen3MTP
Action required
Add Qwen3.5 MTP speculative decoding Build and Test #771: Pull request #351 opened by aleroot
Action required aleroot:Qwen3MTP
aleroot:Qwen3MTP
Action required
Gemma4Text: load conversions that omit redundant KV-shared-layer K/V projections Build and Test #770: Pull request #350 opened by jfranknichols
Action required jfranknichols:gemma4-load-omitted-kv-shared-projections
jfranknichols:gemma4-load-omitted-kv-shared-projections
Action required
Expose ModelTypeRegistry.contains(_:) for pre-load support checks Build and Test #769: Pull request #349 opened by jfranknichols
12m 28s jfranknichols:expose-modeltyperegistry-contains
jfranknichols:expose-modeltyperegistry-contains
12m 28s
Add TranslateGemma support (Gemma 3 translation template) Build and Test #768: Pull request #348 opened by beshkenadze
Action required beshkenadze:upstream-translategemma
beshkenadze:upstream-translategemma
Action required
Gemma 3: chunked prompt prefill, skip lm_head on prompt positions Build and Test #767: Pull request #346 synchronize by beshkenadze
Action required beshkenadze:upstream-gemma3-prefill
beshkenadze:upstream-gemma3-prefill
Action required
Add Hunyuan dense V1 (hunyuan_v1_dense): Hunyuan-MT-7B and Hy-MT2-7B Build and Test #766: Pull request #347 opened by beshkenadze
Action required beshkenadze:upstream-hunyuan
beshkenadze:upstream-hunyuan
Action required
Gemma 3: chunked prompt prefill, skip lm_head on prompt positions Build and Test #765: Pull request #346 opened by beshkenadze
Action required beshkenadze:upstream-gemma3-prefill
beshkenadze:upstream-gemma3-prefill
Action required
Qwen2-VL: implement M-RoPE in the language model (mirror of #239) Build and Test #764: Pull request #345 opened by mnmly
Action required mnmly:qwen2vl-mrope
mnmly:qwen2vl-mrope
Action required
fix MLXVLM prepare(): honor windowSize for chunked prefill (FastVLM / Gemma3 / Qwen2VL / LFM2VL / Pixtral / Mistral3) Build and Test #763: Pull request #344 opened by john-rocky
Action required john-rocky:fix/vlm-prepare-window-size
john-rocky:fix/vlm-prepare-window-size
Action required
Gemma 4: add audio and video multimodal support Build and Test #762: Pull request #343 opened by atlascodesai
Action required atlas-open-sources:gemma4-multimodal-pr-v2
atlas-open-sources:gemma4-multimodal-pr-v2
Action required
Improve Qwen3.5 recurrent cache handling Build and Test #761: Pull request #323 synchronize by aleroot
8m 52s aleroot:qwen3_5perf
aleroot:qwen3_5perf
8m 52s
Improve Qwen3.5 recurrent cache handling Build and Test #760: Pull request #323 synchronize by aleroot
Action required aleroot:qwen3_5perf
aleroot:qwen3_5perf
Action required