nemo-mbridge-perf-activation-recompute skillA
nemo-mbridge-perf-activation-recompute is agent-read markdown (skill) from nvidia/skills: Validate and use selective and full activation recompute in Megatron Bridge to reduce GPU memory usage at the cost of extra compute. Use for activation memory OOMs or regressions involving recompute_granularity, recompute_num_layers, recompute_modules, recompute_method, selective recompute, full recompute, or activation checkpointing..
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.
What the file says
# Activation Recompute Stable docs: @docs/training/activation-recomputation.md Card: @skills/nemo-mbridge-perf-activation-recompute/card.yaml <!-- Guidance refreshed: 2026-08-12. --> Activation recompute (activation checkpointing) trades additional forward work during backward for lower retained-activation memory. The useful checkpoint boundary depends on the model architecture, attention backend, parallelism, and the tensor that actually drives the per-rank peak. ## Quick Decision Guide 1. Confirm the pressure is real allocation, not allocator fragmentation. Compare `max_memory_allocated()` with `max_memory_reserved()` on every rank. 2. Keep an explicit no-recompute control when the workload fits. Under selective granularity, `recompute_modules=[]` is valid and useful for this comparison. 3. Select the first boundary from the architecture and observed peak: - **Standard attention:** `core_attn` is the common first candidate. It is strongest when unfused attention materializes score/probability tensors. With Transformer Engine fused or Flash Attention, compare it against `[]` because those backends already rematerialize attention internals. …
Read the whole file at its exact version.
How to install
mdr add nvidia/skills/nemo-mbridge-perf-activation-recompute@git:20260817.d36ff70mdr add nvidia/skills/nemo-mbridge-perf-activation-recompute@sha256:03379723143dfcd3Pin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.
[](https://markdownregistry.com/a/art_a25z4swu5g26bxlh)
1 badge views in 30 days
Versions
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20260817.d36ff70 latest | 2026-08-17 | d36ff70 | 18,915 B | A | view · diff |
| git:20260616.8491888 | 2026-06-16 | 8491888 | 9,197 B | A | view · diff |
| git:20260602.e695a83 | 2026-06-02 | e695a83 | 10,151 B | A | view · diff |
| git:20260529.7d99fda | 2026-05-29 | 7d99fda | 9,295 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (18915 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
nvidia/skills · 3,415 stars · license Apache-2.0 · pushed 2026-09-23 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_a25z4swu5g26bxlh GET https://markdownregistry.com/api/v1/resolve?ref=nvidia/skills/nemo-mbridge-perf-activation-recompute GET https://markdownregistry.com/api/v1/blob/03379723143dfcd3cce4338c148fecc2aa515649ab85d3624e43a29b80f03e4e
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.