CtrlK
BlogDocsLog inGet started
Tessl Logo

huggingface-accelerate

Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/FSDP/Megatron/DDP. Automatic device placement, mixed precision (FP16/BF16/FP8). Interactive config, single launch command. HuggingFace ecosystem standard.

64

Quality

77%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./08-distributed-training/accelerate/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

72%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

Highly actionable, well-structured content with excellent progressive disclosure via real reference files, held back by some repetition in the workflows and the absence of explicit validation checkpoints.

Suggestions

Tighten Workflow 1 by showing only the diff/added lines instead of repeating the entire original script, and trim the 'Key advantages' recap that restates the description.

Add explicit validation checkpoints to the launch workflow (e.g., verify processes spawned per GPU, confirm loss is decreasing, or check `accelerator.print` only on main process) to raise workflow clarity.

DimensionReasoningScore

Conciseness

Mostly efficient executable code with no basic-concept padding, but Workflow 1 repeats the full original script before the diff and the 'Key advantages' recap mirrors the description, so it could be tightened.

2 / 3

Actionability

Provides copy-paste-ready code and commands throughout — accelerate launch invocations, plugin construction, config JSON, and accelerator.prepare/backward/accumulate usage.

3 / 3

Workflow Clarity

Workflows are clearly sequenced (convert → config → launch) but lack explicit validation checkpoints or feedback loops; 'Common issues' is error guidance rather than an in-workflow checkpoint.

2 / 3

Progressive Disclosure

A concise quick-start overview with well-signaled one-level-deep references to real files (megatron-integration.md, custom-plugins.md, performance.md) for advanced topics, with content appropriately split.

3 / 3

Total

10

/

12

Passed

Description

82%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A specific, well-scoped description with strong capability detail and natural trigger terms, but it omits an explicit 'Use when...' clause, capping its completeness at 2.

Suggestions

Add an explicit 'Use when...' trigger clause (e.g., 'Use when adding distributed/multi-GPU training to a PyTorch script, or when the user mentions DeepSpeed, FSDP, DDP, or mixed precision') to raise completeness.

Consider naming 'accelerate' or 'HuggingFace Accelerate' directly in the description so the trigger term matches what users search for.

DimensionReasoningScore

Specificity

Lists multiple concrete capabilities — '4 lines to add distributed support to any PyTorch script', 'Unified API for DeepSpeed/FSDP/Megatron/DDP', 'Automatic device placement, mixed precision (FP16/BF16/FP8)' — matching the comprehensive-actions anchor.

3 / 3

Completeness

Strongly answers 'what' but lacks any explicit 'Use when...' trigger clause, so per the guidelines completeness is capped at 2 with the 'when' only implied.

2 / 3

Trigger Term Quality

Uses natural practitioner terms — 'distributed training', 'PyTorch', 'DeepSpeed', 'FSDP', 'mixed precision', 'HuggingFace' — that the target audience would actually say.

3 / 3

Distinctiveness Conflict Risk

Occupies a clear niche (HuggingFace Accelerate distributed training) with distinct technology triggers, making conflicts with other skills unlikely.

3 / 3

Total

11

/

12

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository
Orchestra-Research/AI-Research-SKILLs
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.