michaelfeil / infinity

Infinity is a high-throughput, low-latency serving engine for text-embeddings, reranking models, clip, clap and colpali
https://michaelfeil.github.io/infinity/
MIT License
1.46k stars 112 forks source link

cli to set max_seq_length = 2048 #411

Open Liwan-Chen opened 1 month ago

Liwan-Chen commented 1 month ago

Feature request

embedding_model = os.getenv("EMBEDDING_MODEL_FOR_INTENT_TRAINING", "jinaai/jina-embeddings-v2-base-zh") I requested the model jinaai/jina-embeddings-v2-base-zh through infinity. Is there any method to set the model's max_seq_length = 2048?

Motivation

Is there any method to set the model's max_seq_length = 2048?

Your contribution

Is there any method to set the model's max_seq_length = 2048?

michaelfeil commented 1 month ago

https://huggingface.co/jinaai/jina-embeddings-v2-base-zh/discussions/17

can you try infinity_emb v2 --model-id jinaai/jina-embeddings-v2-base-zh --revision refs/pr/17