Releases: huggingface/tei-gaudi
v1.5.3: SynapseAI v1.20.0
You can pull the Docker image associated to this release with:
docker pull ghcr.io/huggingface/tei-gaudi:1.5.3
SynapseAI v1.20.0
The codebase is validated with SynapseAI 1.20.0 and optimum-habana 1.16.0.
Validated Models
Text Embeddings
tei-gaudi currently supports Nomic, BERT, CamemBERT, XLM-RoBERTa models with absolute positions, JinaBERT model with Alibi positions and Mistral, Alibaba GTE and Qwen2 models with Rope positions.
Below are some examples of our validated models:
| Architecture | Pooling | Models |
|---|---|---|
| BERT | Cls/Mean/Last token | |
| BERT | Splade | |
| MPNet | Cls/Mean/Last token | |
| ALBERT | Cls/Mean/Last token | |
| Mistral | Cls/Mean/Last token | |
| GTE | Cls/Mean/Last token | |
| JinaBERT | Cls/Mean/Last token |
Sequence Classification and Re-Ranking
tei-gaudi currently supports CamemBERT, and XLM-RoBERTa Sequence Classification models with absolute positions.
Below are some examples of the currently supported models:
| Task | Model Type | Model ID |
|---|---|---|
| Re-Ranking | XLM-RoBERTa | BAAI/bge-reranker-large |
| Re-Ranking | XLM-RoBERTa | BAAI/bge-reranker-base |
| Sentiment Analysis | RoBERTa | SamLowe/roberta-base-go_emotions |
Full Changelog: huggingface/tgi-gaudi@v1.5.2...v1.5.3
v1.5.2: SynapseAI v1.19.0
You can pull the Docker image associated to this release with:
docker pull ghcr.io/huggingface/tei-gaudi:1.5.2
SynapseAI v1.19.0
The codebase is validated with SynapseAI 1.19.0 and optimum-habana 1.15.0.
Validated Models
Text Embeddings
tei-gaudi currently supports Nomic, BERT, CamemBERT, XLM-RoBERTa models with absolute positions, JinaBERT model with Alibi positions and Mistral, Alibaba GTE and Qwen2 models with Rope positions.
Below are some examples of our validated models:
| Architecture | Pooling | Models |
|---|---|---|
| BERT | Cls/Mean/Last token | |
| BERT | Splade | |
| MPNet | Cls/Mean/Last token | |
| ALBERT | Cls/Mean/Last token | |
| Mistral | Cls/Mean/Last token | |
| GTE | Cls/Mean/Last token | |
| JinaBERT | Cls/Mean/Last token |
Sequence Classification and Re-Ranking
tei-gaudi currently supports CamemBERT, and XLM-RoBERTa Sequence Classification models with absolute positions.
Below are some examples of the currently supported models:
| Task | Model Type | Model ID |
|---|---|---|
| Re-Ranking | XLM-RoBERTa | BAAI/bge-reranker-large |
| Re-Ranking | XLM-RoBERTa | BAAI/bge-reranker-base |
| Sentiment Analysis | RoBERTa | SamLowe/roberta-base-go_emotions |
Full Changelog: huggingface/tgi-gaudi@v1.5.1...v1.5.2
v1.5.1: SynapseAI v1.18.0
You can pull the Docker image associated to this release with:
docker pull ghcr.io/huggingface/tei-gaudi:1.5.1
SynapseAI v1.18.0
The codebase is validated with SynapseAI 1.18.0 and optimum-habana 1.14.1.
Validated Models
| Architecture | Model Type | Models |
|---|---|---|
| BERT | Embedding | |
| MPNet | Embedding | |
| ALBERT | Embedding | |
| Mistral | Embedding | |
| GTE | Embedding | |
| JinaBERT | Embedding | |
| Roberta | Sequence Classification |
Full Changelog: huggingface/tgi-gaudi@v1.5.0...v1.5.1
v1.5.0: SynapseAI v1.17.0
SynapseAI v1.17.0
The codebase is validated with SynapseAI 1.17.0 and optimum-habana 1.13.1.