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

docs: JEPA roadmap — GPU resume re-upload done; remaining items need multi-GPU
CI / Clippy Check (push) Failing after 5s
Details
CI / Format Check (push) Failing after 5s
Details
CI / Build CPU-Only (Explicit) (push) Failing after 6s
Details
CI / Build (ubuntu-latest) (push) Failing after 6s
Details
Documentation / Build User Guide (push) Successful in 4s
Details
CI / Build (macos-latest) (push) Failing after 7s
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
Performance Benchmarks / Run Benchmarks (push) Successful in 25s
Details
Documentation / Build API Documentation (push) Failing after 29s
Details

Browse Source 
Co-Authored-By: Claude Fable 5 <[email protected]>
This commit is contained in:
osobh
2026-07-10 03:59:48 -07:00
co-authored by Claude Fable 5
parent a755627269
commit fce6cef262
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
@@ -152,7 +152,9 @@ All in `crates/training/rtx-transformers/src/ssl/`. 163 tests passing.
**Done (2026-07-10, round 2):** NCCL GPU-direct AllReduce (`nccl` feature; id bootstrap over the TCP rendezvous; verified single-rank on RTX 5060 Ti); k-NN + linear-probe eval wired into the training loop (results in summary/CSV/CLI); GPU-run checkpointing + eval via `context_encoder_cpu_weights()`. **Done (2026-07-10, round 2):** NCCL GPU-direct AllReduce (`nccl` feature; id bootstrap over the TCP rendezvous; verified single-rank on RTX 5060 Ti); k-NN + linear-probe eval wired into the training loop (results in summary/CSV/CLI); GPU-run checkpointing + eval via `context_encoder_cpu_weights()`.
**Next:** GPU weight re-upload on checkpoint resume (GPU resume currently restores only the step counter, with a warning); tensor/pipeline parallel execution (config exists, execution is DP-only — needs a multi-GPU node to build honestly); multi-rank NCCL verification on real multi-GPU hardware. **Done (2026-07-10, round 3):** GPU weight re-upload on checkpoint resume (restore into host copy + push to device buffers; verified live: 20-step train → resume → 10 further steps).
**Next (needs multi-GPU hardware):** tensor/pipeline parallel execution (config exists, execution is DP-only); multi-rank NCCL verification.
## Current State ## Current State