rustyverse/rustytorch
2
0
Fork 0
Code Issues Pull Requests Actions 14 Packages Projects Releases Wiki Activity

docs: JEPA roadmap — 2026-07-10 items done, next tier queued
CI / Build CPU-Only (Explicit) (push) Failing after 6s
Details
CI / Clippy Check (push) Failing after 6s
Details
Performance Benchmarks / Run Benchmarks (push) Failing after 8s
Details
Documentation / Build User Guide (push) Successful in 7s
Details
CI / Build (macos-latest) (push) Failing after 10s
Details
CI / Format Check (push) Failing after 11s
Details
Documentation / Build API Documentation (push) Failing after 14s
Details
CI / Build (ubuntu-latest) (push) Failing after 46s
Details
CI / Test (macos-latest) (push) Has been skipped
Details
CI / Test (ubuntu-latest) (push) Has been skipped
Details
CI / Python Bindings (maturin) (macos-latest) (push) Has been skipped
Details
CI / Python Bindings (maturin) (ubuntu-latest) (push) Has been skipped
Details
CI / WASM Build + Size Check (push) Has been skipped
Details
CI / Distributed Training Tests (push) Has been skipped
Details
CI / CI Success (push) Failing after 0s
Details

Browse Source 
Co-Authored-By: Claude Fable 5 <[email protected]>
This commit is contained in:
osobh
2026-07-10 00:04:52 -07:00
co-authored by Claude Fable 5
parent 19b6581f9c
commit f6b2308381
1 changed files with 3 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CLAUDE.md
+3 -1
View File
@@ -148,7 +148,9 @@ All in `crates/training/rtx-transformers/src/ssl/`. 163 tests passing.
**Done (2026-07-09):** GPU backend wiring — full GPU-resident ViT block in `jepa_gpu.rs` (cuBLAS GEMMs + nvrtc layernorm/GELU/softmax/head-slice kernels, CPU↔GPU parity < 1e-3 verified on RTX 5060 Ti); WebDataset `.tar` filesystem reading (`WebDatasetShard::load`); cluster-plan wiring (`JepaRunConfig::apply_cluster_plan`). **Done (2026-07-09):** GPU backend wiring — full GPU-resident ViT block in `jepa_gpu.rs` (cuBLAS GEMMs + nvrtc layernorm/GELU/softmax/head-slice kernels, CPU↔GPU parity < 1e-3 verified on RTX 5060 Ti); WebDataset `.tar` filesystem reading (`WebDatasetShard::load`); cluster-plan wiring (`JepaRunConfig::apply_cluster_plan`).
**Next:** multi-node AllReduce, training CLI, gzip WebDataset shards, remove the `active_blocks` 2-block cap in jepa_gpu. **Done (2026-07-10):** multi-node TCP AllReduce (`JepaGradSync::tcp`, wired into the runner for world_size > 1, two-rank loopback training tested); `rtx-jepa` training CLI (`crates/tooling/rtx-jepa-cli`: train/bench/plan/validate); gzip WebDataset shards; full-depth GPU and CPU ViT (all bring-up block caps removed, parity verified); `ViTSizeStr::Micro` test size.
**Next:** NCCL/RNCCL GPU-direct AllReduce (TCP backend is CPU-buffer parameter-server), tensor/pipeline parallel execution (config exists, execution is DP-only), linear-probe eval wired into the training loop.
## Current State ## Current State