d-flex flex-column width-full">

fix tied/untied LM-head

1894405

Select commit

Loading

Failed to load commit list.

There was an error while loading. Please reload this page.

Build on: pull_request

Build Wheels on: pull_request

Build wheel (CUDA 12.8)

Build wheel (CUDA 13.0)

Deploy to Modal

Recompute - Determinism test - BF16

Recompute - Determinism test - FP8

Recompute - Full Block - BF16

Recompute - Full Block - FP8

Recompute - Blockwise - BF16

Recompute - Blockwise - FP8

Recompute - Non-linearities - BF16

Recompute - Non-linearities - FP8

Recompute - Offload Opt - BF16

Recompute - Offload Opt - FP8

Recompute - Offload Gradient - BF16

Recompute - Offload Gradient - FP8

Recompute - Chunked attention - BF16

Recompute - Chunked attention - FP8

Test fixed - LLMQ BF16

Test fixed - LLMQ FP8

Test fixed - LLMQ E5M2

Test fixed - Torch BF16 ga=1

Test fixed - Torch BF16 ga=4

Test fixed - Torch FP32

Test fixed - Torch AMP

Test fixed - Torch Chunking

Test fixed - Torch BF16 Custom Matmul

release

Loading