fix: nomic embeddings (#2030)

* fix: allow to configure trust_remote_code

based on: https://github.com/zylon-ai/private-gpt/issues/1893#issuecomment-2118629391

* fix: nomic hf embeddings

poetry.lock

@@ -1151,6 +1151,17 @@ files = [
     {file = "docx2txt-0.8.tar.gz", hash = "sha256:2c06d98d7cfe2d3947e5760a57d924e3ff07745b379c8737723922e7009236e5"},
 ]
 
+[[package]]
+name = "einops"
+version = "0.8.0"
+description = "A new flavour of deep learning operations"
+optional = true
+python-versions = ">=3.8"
+files = [
+    {file = "einops-0.8.0-py3-none-any.whl", hash = "sha256:9572fb63046264a862693b0a87088af3bdc8c068fde03de63453cbbde245465f"},
+    {file = "einops-0.8.0.tar.gz", hash = "sha256:63486517fed345712a8385c100cb279108d9d47e6ae59099b07657e983deae85"},
+]
+
 [[package]]
 name = "email-validator"
 version = "2.2.0"
@@ -6659,7 +6670,7 @@ cffi = ["cffi (>=1.11)"]
 [extras]
 embeddings-azopenai = ["llama-index-embeddings-azure-openai"]
 embeddings-gemini = ["llama-index-embeddings-gemini"]
-embeddings-huggingface = ["llama-index-embeddings-huggingface"]
+embeddings-huggingface = ["einops", "llama-index-embeddings-huggingface"]
 embeddings-ollama = ["llama-index-embeddings-ollama", "ollama"]
 embeddings-openai = ["llama-index-embeddings-openai"]
 embeddings-sagemaker = ["boto3"]
@@ -6682,4 +6693,4 @@ vector-stores-qdrant = ["llama-index-vector-stores-qdrant"]
 [metadata]
 lock-version = "2.0"
 python-versions = ">=3.11,<3.12"
-content-hash = "dce5b88d92bcfa047bf1e4c9fe2dbb9c63eb864d6bbca2340801ac0a2f02a8d4"
+content-hash = "25abbb45bc462dbf056b83c0925b505ad1232484a18e50f07c5e7f517dd84e6f"

private_gpt/components/embedding/embedding_component.py

@@ -31,6 +31,7 @@ class EmbeddingComponent:
                 self.embedding_model = HuggingFaceEmbedding(
                     model_name=settings.huggingface.embedding_hf_model_name,
                     cache_folder=str(models_cache_path),
+                    trust_remote_code=settings.huggingface.trust_remote_code,
                 )
             case "sagemaker":
                 try:

private_gpt/settings/settings.py

@@ -189,6 +189,10 @@ class HuggingFaceSettings(BaseModel):
         None,
         description="Huggingface access token, required to download some models",
     )
+    trust_remote_code: bool = Field(
+        False,
+        description="If set to True, the code from the remote model will be trusted and executed.",
+    )
 
 
 class EmbeddingSettings(BaseModel):

pyproject.toml

@@ -65,6 +65,9 @@ google-generativeai = {version ="^0.5.4", optional = true}
 # Optional Ollama client
 ollama = {version ="^0.3.0", optional = true}
 
+# Optional HF Transformers
+einops = {version = "^0.8.0", optional = true}
+
 [tool.poetry.extras]
 ui = ["gradio", "ffmpy"]
 llms-llama-cpp = ["llama-index-llms-llama-cpp"]
@@ -75,7 +78,7 @@ llms-sagemaker = ["boto3"]
 llms-azopenai = ["llama-index-llms-azure-openai"]
 llms-gemini = ["llama-index-llms-gemini", "google-generativeai"]
 embeddings-ollama = ["llama-index-embeddings-ollama", "ollama"]
-embeddings-huggingface = ["llama-index-embeddings-huggingface"]
+embeddings-huggingface = ["llama-index-embeddings-huggingface", "einops"]
 embeddings-openai = ["llama-index-embeddings-openai"]
 embeddings-sagemaker = ["boto3"]
 embeddings-azopenai = ["llama-index-embeddings-azure-openai"]

settings.yaml

@@ -91,6 +91,9 @@ embedding:
 huggingface:
   embedding_hf_model_name: nomic-ai/nomic-embed-text-v1.5
   access_token: ${HF_TOKEN:}
+  # Warning: Enabling this option will allow the model to download and execute code from the internet.
+  # Nomic AI requires this option to be enabled to use the model, be aware if you are using a different model.
+  trust_remote_code: true
 
 vectorstore:
   database: qdrant