Embedding model for octomind capability / skill auto-activation: ibm-granite/granite-embedding-30m-english (30M params, 6 layers, 384-dim, CLS-pooled, prefix-free, English) fine-tuned on trigger phrases from the octomind-tap capabilities + skills catalog and…
Runs On
What it takes to serve octomind-embed (30M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 0.1 GB | 0.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.0 GB | 0.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.0 GB | 0.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.
Model Card
By Muvon Un Limited, published under apache-2.0, revision 0d12ad93964c.
Embedding model for octomind capability / skill auto-activation: ibm-granite/granite-embedding-30m-english (30M params, 6 layers, 384-dim, CLS-pooled, prefix-free, English) fine-tuned on trigger phrases from the octomind-tap capabilities + skills catalog and blended back into the base as a WiSE-FT model soup, which beats both the base and the raw fine-tune on the runtime gate (mean-of-top-3 cosine + threshold + margin). CachedMultipleNegativesRankingLoss (scale 10) on in-class pairs and positive-aware hard-negative triplets, MatryoshkaLoss over [384, 256, 192, 128, 96], then weight interpolation with the base. - model.safetensors + 1Pooling/ — sentence-transformers layout (fp32).…
Read Muvon Un Limited's full model card
Embedding model for octomind capability / skill auto-activation:
ibm-granite/granite-embedding-30m-english (30M params, 6 layers, 384-dim,
CLS-pooled, prefix-free, English) fine-tuned on trigger phrases from the
octomind-tap capabilities + skills catalog and blended back into the base
as a WiSE-FT model soup, which beats both the base and the raw fine-tune on
the runtime gate (mean-of-top-3 cosine + threshold + margin).
Training: rule-based + LLM paraphrase augmentation, one epoch of
CachedMultipleNegativesRankingLoss (scale 10) on in-class pairs and
positive-aware hard-negative triplets, MatryoshkaLoss over
[384, 256, 192, 128, 96], then weight interpolation with the base.
Files
model.safetensors+1_Pooling/— sentence-transformers layout (fp32).onnx/model.onnx— fp32 graph.onnx/model_quantized.onnx— int8 (weight-only,reduce_range=True); this is what the octomind runtime loads. Pool with CLS as declared in1_Pooling/config.json.
Use
octomind loads onnx:muvon/octomind-embed via octolib's ONNX provider (MODEL_NAME in
octomind/src/embeddings/mod.rs). Runtime thresholds are model-specific and
calibrated against the int8 graph (AUTO_ACTIVATE_THRESHOLD / _MARGIN in
capability.rs, SEMANTIC_* in skill.rs).
Configuration
- Architecture
- RobertaModel
- Context length (tokens)
- 514
- Layers
- 6
- Hidden size
- 384
- Feed-forward size
- 1,536
- Attention heads
- 12
- Vocabulary size
- 50,265
- Model type
- roberta
Identity and Version
- Repository
- muvon/octomind-embed
- Publisher
- Muvon Un Limited
- Task
- Not stated by the source
- Modality
- Other
- Library
- sentence-transformers
- Parameters
- 30M parameters
- Languages
- Not stated by the source
- Revision
- 0d12ad93964c3e1a594509340194311638baecbd
- First published
- 2026-05-12
- Last updated
- 2026-09-18
Files and Weights
30 files, 282.5 MB in total. The weights are 4 files totalling 272.4 MB in bin, onnx, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model.safetensors | Weights | 121.2 MB | 000a1e47356a |
| onnx/model.onnx | Weights | 120.7 MB | 0f3475e1e633 |
| onnx/model_quantized.onnx | Weights | 30.5 MB | ca5ba6a3e4e4 |
| training_args.bin | Weights | 6.3 KB | 6969afc457bc |
| 1_Pooling/config.json | Configuration | 89 B | — |
| config.json | Configuration | 624 B | — |
| config_sentence_transformers.json | Configuration | 284 B | — |
| modules.json | Configuration | 429 B | — |
| onnx/1_Pooling/config.json | Configuration | 89 B | — |
| onnx/calibration.json | Configuration | 11.6 KB | — |
| onnx/config.json | Configuration | 624 B | — |
| onnx/config_sentence_transformers.json | Configuration | 284 B | — |
| onnx/modules.json | Configuration | 429 B | — |
| onnx/sentence_bert_config.json | Configuration | 241 B | — |
| onnx/special_tokens_map.json | Configuration | 957 B | — |
| sentence_bert_config.json | Configuration | 241 B | — |
| special_tokens_map.json | Configuration | 957 B | — |
| README.md | Documentation | 1.5 KB | — |
| eval/Information-Retrieval_evaluation_holdout-ir_results.csv | Other | 1.1 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
| merges.txt | Tokenizer | 456.3 KB | — |
| onnx/merges.txt | Tokenizer | 456.3 KB | — |
| onnx/tokenizer.json | Tokenizer | 3.6 MB | — |
| onnx/tokenizer_config.json | Tokenizer | 1.5 KB | — |
| onnx/vocab.json | Tokenizer | 798.3 KB | — |
| onnx/vocab.txt | Tokenizer | 231.5 KB | — |
| tokenizer.json | Tokenizer | 3.6 MB | — |
| tokenizer_config.json | Tokenizer | 1.5 KB | — |
| vocab.json | Tokenizer | 798.3 KB | — |
| vocab.txt | Tokenizer | 231.5 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 272.4 MB
Released by Muvon Un Limited through its official repository on Hugging Face. Read the license.
Built From
- Derived from ibm-granite/granite-embedding-30m-english
- Quantized from ibm-granite/granite-embedding-30m-english
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 272.4 MB |
| 16-bit | 0.1 GB |
| 8-bit | 0.0 GB |
| 4-bit | 0.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About octomind-embed
How much GPU memory does octomind-embed need?
About 0.1 GB at 16-bit and 0 GB at 4-bit: the weights (30M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run octomind-embed on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use octomind-embed commercially?
Yes. octomind-embed is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
What is octomind-embed's context length?
514 tokens, from the maximum position embeddings in its published configuration.