bechm

It would be great if your lib supported the LongLora technique or other context-expanding ones.

Missing node update in multithreading mode of cluster tree builder

1

The script https://github.com/parthsarthi03/raptor/blob/2e3e83e5c4aa6a9b5f2d8359f5b71a9159c20845/raptor/cluster_tree_builder.py#L114C1-L126C49 ```if use_multithreading: with ThreadPoolExecutor() as executor: for cluster in clusters: executor.submit( process_cluster, cluster, new_level_nodes, next_node_index, summarization_length, lock, ) next_node_index += 1 executor.shutdown(wait=True) ``` runs the risk of...

TGI has released the 2.3.1 version. This is important as it supports inference with Llama3.2-vision. Please consider upgrade the image so we can use the model on Sagemaker.

Deploy error for Llama-3.2-vision-11B: "Sharded is not supported for AutoModel"

4

### System Info Hi Team, When deploying the model on AWS with `huggingface-pytorch-tgi-inference:2.3.0-tgi2.2.0`, I got the above error. Could you tell me when can TGI provide the new image? Is...