skilld
All skills
huggingface avatar

/train-sentence-transformers

@0201949 official
by Hugging Facehuggingface/skills11k stars
753

Train or fine-tune sentence-transformers models across `SentenceTransformer` (bi-encoder, dense or static embedding model for retrieval, similarity, clustering, classification, paraphrase mining, dedup, multimodal), `CrossEncoder` (reranker, pair scoring for two-stage retrieval / pair classification), `SparseEncoder` (SPLADE, sparse embedding model for learned-sparse retrieval), and `MultiVectorEncoder` (ColBERT / late-interaction, per-token embeddings scored with MaxSim). Covers loss selection, hard-negative mining, evaluators, distillation, LoRA, Matryoshka, and Hugging Face Hub publishing. Use for any sentence-transformers training task.

Ask your Agent

Use this Skill: https://skilld.dev/gh/huggingface/skills/train-sentence-transformers

This session only. Nothing lands on disk.

SKILL.md

≈169 tokens always: the name and description. ≈2.4k when used: this file. ≈33k more on demand in 16 files.

Train a sentence-transformers Model

This SKILL.md is a router, not a manual. It tells you which references and example scripts to load for your task. The actual content (recommended losses, evaluators, training-script structure, model selection, training-arg knobs, troubleshooting) lives in references/ and scripts/.

Do not synthesize a training script from this file alone. Open the per-type production template (scripts/train_<type>_example.py) and copy it as your starting point. The templates contain load-bearing scaffolding (autocast helper, model-card class, logger silencing list, force=True, seed, TF32, version-compatible imports, named-evaluator metric handling) that prior agent runs have repeatedly missed when rolling their own from a synthesized snippet.

1. Identify the model type

Tag Class What it does When to pick
[SentenceTransformer] SentenceTransformer (bi-encoder) Maps each input to a fixed-dim dense vector Retrieval, similarity, clustering, classification, paraphrase mining, dedup
[CrossEncoder] CrossEncoder (reranker) Scores (query, passage) pairs jointly Two-stage retrieval (rerank top-100 from bi-encoder), pair classification
[SparseEncoder] SparseEncoder (SPLADE) Sparse vectors over the vocabulary Learned-sparse retrieval, inverted-index backends (Elasticsearch / OpenSearch / Lucene)
[MultiVectorEncoder] MultiVectorEncoder (ColBERT) One embedding per token, scored with MaxSim Late-interaction retrieval, recall gains over bi-encoders at higher storage cost, multimodal (ColPali / ColQwen2)

Tiebreakers when the request is ambiguous: "embedding model" / "vector search" / "similarity" → [SentenceTransformer]. "rerank" / "ranker" / "two-stage" → [CrossEncoder]. "SPLADE" / "sparse" / "inverted index" → [SparseEncoder]. "ColBERT" / "late interaction" / "multi-vector" / "MaxSim" / "ColPali" / "ColQwen" → [MultiVectorEncoder]. If still unclear, ask.

2. Required reading

Read these in full before writing any code. Do not triage by perceived relevance.

Per-type: always required

[SentenceTransformer]

  • references/losses_sentence_transformer.md: loss-to-data-shape mapping, BatchSamplers.NO_DUPLICATES requirement for MNRL-family, Cached* ↔ gradient_checkpointing incompatibility.
  • references/evaluators_sentence_transformer.md: evaluator-to-task mapping, metric_for_best_model key construction (named vs unnamed), per-evaluator primary_metric values.
  • references/model_architectures.md: encoder vs decoder vs static vs Router pipelines, pooling rules (mean / cls / lasttoken), auto-mean-pooling behavior for fresh-start MLM bases.
  • scripts/train_sentence_transformer_example.py: production template. Copy this as your starting point.

[CrossEncoder]

  • references/losses_cross_encoder.md: pointwise / pairwise / listwise / distillation, pos_weight derivation, activation_fn=Identity() mandatory for non-BCE losses (silent eval-rank collapse otherwise).
  • references/evaluators_cross_encoder.md: CrossEncoderRerankingEvaluator recipe, named-evaluator key format eval_{name}_{primary_metric}.
  • scripts/train_cross_encoder_example.py: production template. Copy this as your starting point.

[SparseEncoder]

  • references/losses_sparse_encoder.md: SpladeLoss wrapper requirement, FLOPS regularizer weights, smoke-test active-dim ramp behavior.
  • references/evaluators_sparse_encoder.md: SparseNanoBEIREvaluator (English-only) and the in-domain alternative, eval_{name}_{primary_metric} key format.
  • scripts/train_sparse_encoder_example.py: production template. Copy this as your starting point.

[MultiVectorEncoder]

  • references/losses_multi_vector_encoder.md: MaxSim scoring, scale choice per scoring mode (scale=1.0 for MaxSim, roughly the average query length for MeanMaxSim), MNRL / CachedMNRL / MarginMSE / DistillKLDiv, XTR-vs-ColBERT scoring, CachedMNRL ↔ gradient_checkpointing incompatibility.
  • references/evaluators_multi_vector_encoder.md: MultiVectorNanoBEIREvaluator (English-only) and the in-domain alternative, eval_NanoBEIR_mean_maxsim_ndcg@10 key format, distillation-eval spearman variant.
  • scripts/train_multi_vector_encoder_example.py: production template. Copy this as your starting point.

Cross-cutting: always required (regardless of task)

  • references/training_args.md: TrainingArguments knobs, precision rules (load fp32 + autocast bf16/fp16, never torch_dtype=bfloat16), warmup_steps (float) vs deprecated warmup_ratio, save_steps must be a multiple of eval_steps for load_best_model_at_end, schedulers, HPO, tracker, resume, hub-push variants.
  • references/dataset_formats.md: column-matching rules (label name auto-detection, column-order-not-name), reshaping recipes, hard-negative mining options.
  • references/base_model_selection.md: discovery commands, per-type model namespaces, ModernBERT-family max_seq_length=8192 trap, datasets >= 4 script-loader rejection, non-English starting-point shortcuts.
  • references/troubleshooting.md: symptom-indexed failure recipes. Skim the section headings on every run, even a healthy one. The "Metrics don't improve" and "Hub push fails" entries cover bugs that bite frequently and are cheaper to recognize before they fire than to debug after.

Cross-cutting: load when applicable

  • references/hardware_guide.md: VRAM sizing, multi-GPU, FSDP / DeepSpeed, HF Jobs flavors. Required for >24GB models, multi-GPU, or HF Jobs runs.
  • references/hf_jobs_execution.md: required when running on HF Jobs.
  • references/prompts_and_instructions.md: required when using prompt-tuned bases (E5, BGE, GTE, Qwen3-Embedding, Instructor, Nomic, etc.) or adding query: / passage: style prefixes.

Variant scripts (open when the task matches)

  • [SentenceTransformer] scripts/train_sentence_transformer_<matryoshka|multi_dataset|with_lora|distillation|make_multilingual|static_embedding>_example.py.
  • [CrossEncoder] scripts/train_cross_encoder_<distillation|listwise>_example.py.
  • [SparseEncoder] scripts/train_sparse_encoder_distillation_example.py.
  • Hard-negative mining CLI: scripts/mine_hard_negatives.py.

3. Defaults

Override only if the user specifies otherwise:

  • Local execution. Pitch HF Jobs only if local hardware can't fit the job.
  • Single run. After it completes, propose experimentation if the user would benefit (weak/marginal verdict, "see how high you can push it" framing, etc.). Iteration rules in references/training_args.md (Experimentation section).
  • Public Hub push at end-of-run, wrapped in try-except. On HF Jobs (ephemeral env) ALSO enable in-trainer push (push_to_hub=True + hub_strategy="every_save"). Details in references/hf_jobs_execution.md.

4. Constraints the produced script must satisfy

These are non-negotiable contracts. Implementation lives in the production templates and references. Do not reinvent.

  • Capture the pre-training evaluator score as baseline_eval before trainer.train().
  • Emit a single end-of-run line: VERDICT: WIN|MARGINAL|REGRESSION | score=... | baseline=... | delta=.... A monitor scrapes for this.
  • Silence httpx, httpcore, huggingface_hub, urllib3, filelock, fsspec to WARNING (otherwise HF download URLs flood the agent's context).
  • Tee logs to logs/{RUN_NAME}.log.
  • End with model.push_to_hub(...) wrapped in try/except.
  • Smoke-test before any long run (max_steps=1 + tiny dataset slice). The production templates show one common pattern (SMOKE_TEST env var).
  • [CrossEncoder] Include EarlyStoppingCallback(patience>=3). CE rerankers often peak mid-training and regress.
  • [SparseEncoder] Log query_active_dims / corpus_active_dims on the verdict line. High nDCG with collapsed sparsity is not a win. The keys come back name-prefixed (e.g. ..._query_active_dims). Use suffix matching to pluck them. See the SPARSE production template for the exact pattern.
  • [MultiVectorEncoder] Match scale to the scoring mode on any MNRL-family loss: near 1.0 for unnormalized MaxSim (do not copy scale=20.0 from bi-encoder MNRL), roughly the average query length with length-normalized MeanMaxSim, since each score is divided by its query's token count. XTRScores is a train-only similarity_fct: the evaluators reject it, so evaluation always scores with MaxSim, including for XTR-trained models.

5. Workflow

  1. Identify the model type (§1). Ask if ambiguous.
  2. Load the §2 required-reading files for that type.
  3. Open scripts/train_<type>_example.py and copy it as your starting point.
  4. Replace MODEL_NAME, DATASET_NAME, RUN_NAME, the loss, and the evaluator with the user's task. Cross-check loss/data-shape match against references/losses_<type>.md. Cross-check the metric_for_best_model key against references/evaluators_<type>.md (named evaluators format the key as eval_{name}_{primary_metric}).
  5. Smoke-test (max_steps=1).
  6. Run.
  7. After the run, append to logs/experiments.md and propose iteration if the verdict is weak/marginal.

Prerequisites

pip install "sentence-transformers[train]>=5.0"        # add [train,image] / [audio] / [video] for [SentenceTransformer] multimodal
                                                       # [MultiVectorEncoder] requires >=6.0
pip install trackio                                    # optional tracker (or wandb / tensorboard / mlflow)
hf auth login                                          # or set HF_TOKEN with write scope (for Hub push)

GPU strongly recommended. CPU works only for demos and [SentenceTransformer] StaticEmbedding.

Source: SKILL.md on GitHub

Third-party checks

1 warning16d3 checks · Risk SAFE
  • Gen Agent Trust Hub16d

    This skill provides a comprehensive environment for training sentence-transformers models, including production-ready scripts and detailed documentation. It involves standard practices such as downloading packages from official registries and processing datasets from remote sources.

  • Socket16d

    No alerts

  • Snyk16d

    Risk: MEDIUM · 1 issue

Provenance

Signed by skilld at 0201949. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last week.

Activeupdated last month

Topics

  • sentence-transformers
  • training
  • fine-tuning
  • embeddings
  • retrieval
  • cross-encoder
  • sparse-encoder
  • ner
  • classification

README badge

README badge for huggingface/skills/train-sentence-transformers

What it does

Trains or fine-tunes sentence-transformers models across SentenceTransformer (bi-encoder for dense embeddings), CrossEncoder (reranker for pair scoring), and SparseEncoder (SPLADE for sparse vectors). Covers loss selection, hard-negative mining, evaluators, distillation, LoRA, and Hub publishing. Use this skill for any sentence-transformers training task.

Generated from the current SKILL.md.

Frequently asked

Does this skill cover all three model types (SentenceTransformer, CrossEncoder, SparseEncoder)?
Yes. The skill routes you to type-specific references and production templates. Use section 1 to identify which model type matches your task, then load the corresponding references and example script.
Can I use this skill to fine-tune models with LoRA, distillation, or Matryoshka?
Yes. The skill includes variant scripts for `train_sentence_transformer_with_lora_example.py`, `train_sentence_transformer_distillation_example.py`, and `train_sentence_transformer_matryoshka_example.py`, plus distillation variants for CrossEncoder and SparseEncoder.
Do I need to write my own training script or can I copy from the templates?
Copy from the production templates (`scripts/train_<type>_example.py`). The skill explicitly states not to synthesize from the routing file alone; templates contain load-bearing scaffolding (autocast helpers, seed handling, version-compatible imports, required callbacks) that prior runs have missed when rolling their own.
What if my task involves hard-negative mining or training on multiple datasets?
The skill includes `scripts/mine_hard_negatives.py` for hard-negative mining and a `train_sentence_transformer_multi_dataset_example.py` variant. Check section 2 (Variant scripts) and `references/dataset_formats.md` for reshaping recipes.
Does this work with multimodal models or non-English languages?
For multimodal: install `sentence-transformers[train,image]` or add audio/video extras. For non-English: the skill references `references/base_model_selection.md` (non-English shortcuts) and `references/prompts_and_instructions.md` for prompt-tuned bases (E5, BGE, Qwen3-Embedding, etc.).

Generated from the current SKILL.md. These answers refresh after source changes.