Skip to content

Actions: ashhart/TensorFold

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
413 workflow runs
413 workflow runs

Filter by Workflow

⌥ + click/return to exclude

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
Nemotron: fix SSM backward thread limit on M1 Zig format #189: Pull request #715 opened by mrv777
Action required mrv777:fix/nemotron-training-m1
mrv777:fix/nemotron-training-m1
Action required
Nemotron: fix SSM backward thread limit on M1 No GPU #191: Pull request #715 opened by mrv777
Action required mrv777:fix/nemotron-training-m1
mrv777:fix/nemotron-training-m1
Action required
fix: replace whole subjects in Living Weights neighbour questions Zig format #188: Pull request #714 opened by cwschroeder
Action required cwschroeder:fix/living-subject-neighbours
cwschroeder:fix/living-subject-neighbours
Action required
fix: replace whole subjects in Living Weights neighbour questions No GPU #190: Pull request #714 opened by cwschroeder
Action required cwschroeder:fix/living-subject-neighbours
cwschroeder:fix/living-subject-neighbours
Action required
perf(qwen27): a drafted round overlaps its cache commit with the next draft, and the drafter's attention and top-k use more of the GPU Zig format #187: Pull request #713 opened by jkuepker
40s jkuepker:qwen27-drafted-round
jkuepker:qwen27-drafted-round
40s
perf(qwen27): a drafted round overlaps its cache commit with the next draft, and the drafter's attention and top-k use more of the GPU No GPU #189: Pull request #713 opened by jkuepker
2m 38s jkuepker:qwen27-drafted-round
jkuepker:qwen27-drafted-round
2m 38s
perf(qwen27): decode attention without the tail kernel, the gate dispatch and the idle simdgroups Zig format #186: Pull request #712 opened by jkuepker
35s jkuepker:qwen27-decode-attention-chain
jkuepker:qwen27-decode-attention-chain
35s
perf(qwen27): decode attention without the tail kernel, the gate dispatch and the idle simdgroups No GPU #188: Pull request #712 opened by jkuepker
1m 42s jkuepker:qwen27-decode-attention-chain
jkuepker:qwen27-decode-attention-chain
1m 42s
perf(qwen27): an accepted DeltaNet chain scans its state in place, no state publish copy No GPU #187: Pull request #711 opened by jkuepker
2m 0s jkuepker:qwen27-gdn-chain-inplace
jkuepker:qwen27-gdn-chain-inplace
2m 0s
perf(qwen27): an accepted DeltaNet chain scans its state in place, no state publish copy Zig format #185: Pull request #711 opened by jkuepker
21s jkuepker:qwen27-gdn-chain-inplace
jkuepker:qwen27-gdn-chain-inplace
21s
Add tensorfold merge: fold a PEFT LoRA into a quantized checkpoint No GPU #186: Pull request #710 opened by philip-pentatonic
2m 42s Pentatonic-Ltd:lora-merge-quantize
Pentatonic-Ltd:lora-merge-quantize
2m 42s
Add tensorfold merge: fold a PEFT LoRA into a quantized checkpoint Zig format #184: Pull request #710 opened by philip-pentatonic
55s Pentatonic-Ltd:lora-merge-quantize
Pentatonic-Ltd:lora-merge-quantize
55s
Add MLX's affine 4-bit quantizer and a check against shipped tensors Zig format #183: Pull request #709 opened by philip-pentatonic
24s Pentatonic-Ltd:mlx-affine-quantizer
Pentatonic-Ltd:mlx-affine-quantizer
24s
Add MLX's affine 4-bit quantizer and a check against shipped tensors No GPU #185: Pull request #709 opened by philip-pentatonic
2m 45s Pentatonic-Ltd:mlx-affine-quantizer
Pentatonic-Ltd:mlx-affine-quantizer
2m 45s
GLM: fix KDA and router thread limits on M1 Zig format #182: Pull request #708 opened by mrv777
Action required mrv777:fix/glm-kda-router-m1
mrv777:fix/glm-kda-router-m1
Action required
GLM: fix KDA and router thread limits on M1 No GPU #184: Pull request #708 opened by mrv777
Action required mrv777:fix/glm-kda-router-m1
mrv777:fix/glm-kda-router-m1
Action required
DeepSeek-V4.1: performance hooks and host checks No GPU #183: Pull request #707 opened by jayleaton
5m 16s jayleaton:stack/13a-performance-host
jayleaton:stack/13a-performance-host
5m 16s
DeepSeek-V4.1: performance hooks and host checks Zig format #181: Pull request #707 opened by jayleaton
21s jayleaton:stack/13a-performance-host
jayleaton:stack/13a-performance-host
21s
Name the native Flash Next checkpoint in its recipe Zig format #180: Pull request #705 opened by cwschroeder
Action required cwschroeder:flashnext-recipe-native
cwschroeder:flashnext-recipe-native
Action required
Name the native Flash Next checkpoint in its recipe No GPU #182: Pull request #705 opened by cwschroeder
Action required cwschroeder:flashnext-recipe-native
cwschroeder:flashnext-recipe-native
Action required
GLM: fix sparse attention thread limit on M1 No GPU #181: Pull request #704 opened by mrv777
Action required mrv777:fix/glm-sparse-attention-m1
mrv777:fix/glm-sparse-attention-m1
Action required
GLM: fix sparse attention thread limit on M1 Zig format #179: Pull request #704 opened by mrv777
Action required mrv777:fix/glm-sparse-attention-m1
mrv777:fix/glm-sparse-attention-m1
Action required
DeepSeek-V4.1: served engine, model boot and CLI No GPU #180: Pull request #703 opened by jayleaton
4m 10s jayleaton:stack/12b-serving-engine
jayleaton:stack/12b-serving-engine
4m 10s
DeepSeek-V4.1: served engine, model boot and CLI Zig format #178: Pull request #703 opened by jayleaton
24s jayleaton:stack/12b-serving-engine
jayleaton:stack/12b-serving-engine
24s
DeepSeek-V4.1: serving host, release tokenizer and HTTP family adapter Zig format #177: Pull request #702 opened by jayleaton
30s jayleaton:stack/12a-serving-host
jayleaton:stack/12a-serving-host
30s