Running out of memory when running ingestion pipeline.

[rahulkrprajapati]

When running my ingestion pipeline like:

node_parser = HierarchicalNodeParser.from_defaults(chunk_sizes=[2048, 1024, 256])
pipeline = IngestionPipeline(
                transformations=[
                   node_parser,
                    HuggingFaceEmbedding(
                        model_name='nomic-ai/nomic-embed-text-v1'
                    ),
                ]
            )
           

Traceback (most recent call last):
  File "/Users/rahulprajapati/VSCode/ProtoSearchBackend/DocCore/picore.py", line 422, in process_streamlit_file
    leaf_nodes = get_leaf_nodes(nodes)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/llama_index/core/ingestion/pipeline.py", line 446, in run
    nodes = run_transformations(
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/llama_index/core/ingestion/pipeline.py", line 86, in run_transformations
    nodes = transform(nodes, **kwargs)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/llama_index/core/base/embeddings/base.py", line 334, in __call__
    embeddings = self.get_text_embedding_batch(
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/llama_index/core/base/embeddings/base.py", line 255, in get_text_embedding_batch
    embeddings = self._get_text_embeddings(cur_batch)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/llama_index/embeddings/huggingface/base.py", line 202, in _get_text_embeddings
    return self._embed(texts)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/llama_index/embeddings/huggingface/base.py", line 161, in _embed
    model_output = self._model(**encoded_input)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
    return forward_call(*args, **kwargs)
  File "/Users/rahulprajapati/.cache/huggingface/modules/transformers_modules/nomic-ai/nomic-embed-text-v1/289f532e14dbbbd5a04753fa58739e9ba766f3c7/modeling_hf_nomic_bert.py", line 1070, in forward
    sequence_output = self.encoder(hidden_states, attention_mask=attention_mask, return_dict=return_dict)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
    return forward_call(*args, **kwargs)
  File "/Users/rahulprajapati/.cache/huggingface/modules/transformers_modules/nomic-ai/nomic-embed-text-v1/289f532e14dbbbd5a04753fa58739e9ba766f3c7/modeling_hf_nomic_bert.py", line 957, in forward
    hidden_states, hidden_states2, residual = layer(
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
    return forward_call(*args, **kwargs)
  File "/Users/rahulprajapati/.cache/huggingface/modules/transformers_modules/nomic-ai/nomic-embed-text-v1/289f532e14dbbbd5a04753fa58739e9ba766f3c7/modeling_hf_nomic_bert.py", line 890, in forward
    attn_outputs = self.attn(
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/Users/rahulprajapati/anaconda3/envs/supabaseProto/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
    return forward_call(*args, **kwargs)
  File "/Users/rahulprajapati/.cache/huggingface/modules/transformers_modules/nomic-ai/nomic-embed-text-v1/289f532e14dbbbd5a04753fa58739e9ba766f3c7/modeling_hf_nomic_bert.py", line 793, in forward
    attention_scores = torch.matmul(query, key.transpose(-1, -2)) / self.norm_factor
RuntimeError: MPS backend out of memory (MPS allocated: 16.80 GB, other allocations: 966.88 MB, max allowed: 18.13 GB). Tried to allocate 720.99 MB on private pool. Use PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.0 to disable upper limit for memory allocations (may cause system failure).

[rahulkrprajapati]

@dosu-bot check the stack trace