Model catalog
Every model candding ships, generated from the registry, with its pooling, length limit and backend verification status.
The table is generated from the registry by just docs-models; the same data drives the README table and the candding list-models --json command.
| Model | Family | Dim | Pooling | Max length | CPU | Metal | CUDA | License |
|---|---|---|---|---|---|---|---|---|
| BAAI/bge-small-en-v1.5 | BERT | 384 | cls | 512 | verified | verified | untested | MIT |
| BAAI/bge-base-en-v1.5 | BERT | 768 | cls | 512 | verified | verified | untested | MIT |
| BAAI/bge-large-en-v1.5 | BERT | 1024 | cls | 512 | verified | verified | untested | MIT |
| BAAI/bge-small-zh-v1.5 | BERT | 512 | cls | 512 | verified | verified | untested | MIT |
| sentence-transformers/all-MiniLM-L6-v2 | BERT | 384 | mean | 256 | verified | verified | untested | Apache-2.0 |
| thenlper/gte-base | BERT | 768 | mean | 512 | verified | verified | untested | MIT |
| thenlper/gte-large | BERT | 1024 | mean | 512 | verified | verified | untested | MIT |
| mixedbread-ai/mxbai-embed-large-v1 | BERT | 1024 | cls | 512 | verified | verified | untested | Apache-2.0 |
| Snowflake/snowflake-arctic-embed-xs | BERT | 384 | cls | 512 | verified | verified | untested | Apache-2.0 |
| Snowflake/snowflake-arctic-embed-s | BERT | 384 | cls | 512 | verified | verified | untested | Apache-2.0 |
| Snowflake/snowflake-arctic-embed-m | BERT | 768 | cls | 512 | verified | verified | untested | Apache-2.0 |
| Snowflake/snowflake-arctic-embed-l | BERT | 1024 | cls | 512 | verified | verified | untested | Apache-2.0 |
| sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 | BERT | 384 | mean | 128 | verified | verified | untested | Apache-2.0 |
Reading the columns
- Pooling is how per-token states become one vector:
cls,meanorlast_token; see pooling and templates. - Max length is the default truncation, taken from the repository's sentence-transformers configuration; the builder can raise it up to the model's position limit.
- CPU, Metal, CUDA show verified when the golden suite (CPU) or the parity test (accelerators) passed, pending when a run is planned, and untested when nobody has run it yet; see testing.
- License is the model's license on Hugging Face; candding itself is Apache-2.0 OR MIT.
How a model gets verified
A model is added to the registry with its pooling, templates and default length, a reference is generated with sentence-transformers in F32 on CPU, and the ten golden tests in candding/tests/golden_bert.rs must pass at cosine ≥ 0.9999 and max-abs ≤ 1e-4 before the entry lands. Metal and CUDA become verified when the parity test agrees with CPU at cosine ≥ 0.999 on every fixture input.
Comparison
Where candding sits next to fastembed, fastembed-rs, EmbedAnything, text-embeddings-inference, candle-transformers and hosted embedding APIs.
BERT
The BERT encoder family behind bge, MiniLM, gte, mxbai and Snowflake Arctic embeddings, and the details that make its vectors match the reference.