Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat: Added nomic-embed-text-v1 support + formatting changes + import fixes #118

Merged
merged 3 commits into from
Feb 19, 2024
Merged

feat: Added nomic-embed-text-v1 support + formatting changes + import fixes #118

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
6393276
feat: Added nomic-embed-text-v1 support
Anush008 Feb 13, 2024
86cc4b1
chore: xenova/nomic-embed-text-v1 -> nomic-ai/nomic-embed-text-v1
Anush008 Feb 14, 2024
baf45e4
Merge branch 'main' into nomic-text-embed
NirantK Feb 19, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
# ⚡️ What is FastEmbed?

FastEmbed is a lightweight, fast, Python library built for embedding generation. We [support popular text models](https://qdrant.github.io/fastembed/examples/Supported_Models/). Please [open a Github issue](https://github.com/qdrant/fastembed/issues/new) if you want us to add a new model.
FastEmbed is a lightweight, fast, Python library built for embedding generation. We [support popular text models](https://qdrant.github.io/fastembed/examples/Supported_Models/). Please [open a GitHub issue](https://github.com/qdrant/fastembed/issues/new) if you want us to add a new model.

The default text embedding (`TextEmbedding`) model is Flag Embedding, the top model in the [MTEB](https://huggingface.co/spaces/mteb/leaderboard) leaderboard. It supports "query" and "passage" prefixes for the input text. Here is an example for [Retrieval Embedding Generation](https://qdrant.github.io/fastembed/examples/Retrieval_with_FastEmbed/) and how to use [FastEmbed with Qdrant](https://qdrant.github.io/fastembed/examples/Usage_With_Qdrant/).

Expand Down Expand Up @@ -48,7 +48,7 @@ Installation with Qdrant Client in Python:
pip install qdrant-client[fastembed]
```

Might have to use ```pip install 'qdrant-client[fastembed]'``` on zsh.
You might have to use ```pip install 'qdrant-client[fastembed]'``` on zsh.

```python
from qdrant_client import QdrantClient
Expand Down
50 changes: 31 additions & 19 deletions docs/examples/Supported_Models.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -110,30 +110,38 @@
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>nomic-ai/nomic-embed-text-v1</td>\n",
" <td>768</td>\n",
" <td>8192 context length english model</td>\n",
" <td>0.54</td>\n",
" <td>{'hf': 'xenova/nomic-embed-text-v1'}</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>intfloat/multilingual-e5-large</td>\n",
" <td>1024</td>\n",
" <td>Multilingual model, e5-large. Recommend using this model for non-English languages</td>\n",
" <td>2.24</td>\n",
" <td>{'url': 'https://storage.googleapis.com/qdrant-fastembed/fast-multilingual-e5-large.tar.gz', 'hf': 'qdrant/multilingual-e5-large-onnx'}</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <th>10</th>\n",
" <td>sentence-transformers/paraphrase-multilingual-mpnet-base-v2</td>\n",
" <td>768</td>\n",
" <td>Sentence-transformers model for tasks like clustering or semantic search</td>\n",
" <td>1.11</td>\n",
" <td>{'hf': 'xenova/paraphrase-multilingual-mpnet-base-v2'}</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <th>11</th>\n",
" <td>jinaai/jina-embeddings-v2-base-en</td>\n",
" <td>768</td>\n",
" <td>English embedding model supporting 8192 sequence length</td>\n",
" <td>0.55</td>\n",
" <td>{'hf': 'xenova/jina-embeddings-v2-base-en'}</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <th>12</th>\n",
" <td>jinaai/jina-embeddings-v2-small-en</td>\n",
" <td>512</td>\n",
" <td>English embedding model supporting 8192 sequence length</td>\n",
Expand All @@ -154,10 +162,11 @@
"5 BAAI/bge-small-en-v1.5 384 \n",
"6 BAAI/bge-small-zh-v1.5 512 \n",
"7 sentence-transformers/all-MiniLM-L6-v2 384 \n",
"8 intfloat/multilingual-e5-large 1024 \n",
"9 sentence-transformers/paraphrase-multilingual-mpnet-base-v2 768 \n",
"10 jinaai/jina-embeddings-v2-base-en 768 \n",
"11 jinaai/jina-embeddings-v2-small-en 512 \n",
"8 nomic-ai/nomic-embed-text-v1 768 \n",
"9 intfloat/multilingual-e5-large 1024 \n",
"10 sentence-transformers/paraphrase-multilingual-mpnet-base-v2 768 \n",
"11 jinaai/jina-embeddings-v2-base-en 768 \n",
"12 jinaai/jina-embeddings-v2-small-en 512 \n",
"\n",
" description \\\n",
"0 Base English model \n",
Expand All @@ -168,10 +177,11 @@
"5 Fast and Default English model \n",
"6 Fast and recommended Chinese model \n",
"7 Sentence Transformer model, MiniLM-L6-v2 \n",
"8 Multilingual model, e5-large. Recommend using this model for non-English languages \n",
"9 Sentence-transformers model for tasks like clustering or semantic search \n",
"10 English embedding model supporting 8192 sequence length \n",
"8 8192 context length english model \n",
"9 Multilingual model, e5-large. Recommend using this model for non-English languages \n",
"10 Sentence-transformers model for tasks like clustering or semantic search \n",
"11 English embedding model supporting 8192 sequence length \n",
"12 English embedding model supporting 8192 sequence length \n",
"\n",
" size_in_GB \\\n",
"0 0.50 \n",
Expand All @@ -182,10 +192,11 @@
"5 0.13 \n",
"6 0.10 \n",
"7 0.09 \n",
"8 2.24 \n",
"9 1.11 \n",
"10 0.55 \n",
"11 0.13 \n",
"8 0.54 \n",
"9 2.24 \n",
"10 1.11 \n",
"11 0.55 \n",
"12 0.13 \n",
"\n",
" sources \n",
"0 {'url': 'https://storage.googleapis.com/qdrant-fastembed/fast-bge-base-en.tar.gz'} \n",
Expand All @@ -196,10 +207,11 @@
"5 {'url': 'https://storage.googleapis.com/qdrant-fastembed/fast-bge-small-en-v1.5.tar.gz', 'hf': 'qdrant/bge-small-en-v1.5-onnx-q'} \n",
"6 {'url': 'https://storage.googleapis.com/qdrant-fastembed/fast-bge-small-zh-v1.5.tar.gz'} \n",
"7 {'url': 'https://storage.googleapis.com/qdrant-fastembed/sentence-transformers-all-MiniLM-L6-v2.tar.gz', 'hf': 'qdrant/all-MiniLM-L6-v2-onnx'} \n",
"8 {'url': 'https://storage.googleapis.com/qdrant-fastembed/fast-multilingual-e5-large.tar.gz', 'hf': 'qdrant/multilingual-e5-large-onnx'} \n",
"9 {'hf': 'xenova/paraphrase-multilingual-mpnet-base-v2'} \n",
"10 {'hf': 'xenova/jina-embeddings-v2-base-en'} \n",
"11 {'hf': 'xenova/jina-embeddings-v2-small-en'} "
"8 {'hf': 'xenova/nomic-embed-text-v1'} \n",
"9 {'url': 'https://storage.googleapis.com/qdrant-fastembed/fast-multilingual-e5-large.tar.gz', 'hf': 'qdrant/multilingual-e5-large-onnx'} \n",
"10 {'hf': 'xenova/paraphrase-multilingual-mpnet-base-v2'} \n",
"11 {'hf': 'xenova/jina-embeddings-v2-base-en'} \n",
"12 {'hf': 'xenova/jina-embeddings-v2-small-en'} "
]
},
"execution_count": 4,
Expand Down Expand Up @@ -232,7 +244,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.11.5"
"version": "3.11.7"
},
"orig_nbformat": 4
},
Expand Down
4 changes: 2 additions & 2 deletions fastembed/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,3 @@
from .embedding import TextEmbedding
from fastembed.text.text_embedding import TextEmbedding

__all__ = [TextEmbedding]
__all__ = ["TextEmbedding"]
2 changes: 1 addition & 1 deletion fastembed/common/models.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,7 @@ def load_tokenizer(model_dir: Path, max_length: int = 512) -> Tokenizer:

tokenizer = Tokenizer.from_file(str(tokenizer_path))
tokenizer.enable_truncation(max_length=min(tokenizer_config["model_max_length"], max_length))
tokenizer.enable_padding(pad_id=config["pad_token_id"], pad_token=tokenizer_config["pad_token"])
tokenizer.enable_padding(pad_id=config.get("pad_token_id", 0), pad_token=tokenizer_config["pad_token"])

for token in tokens_map.values():
if isinstance(token, str):
Expand Down
8 changes: 4 additions & 4 deletions fastembed/common/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,10 +24,10 @@ def define_cache_dir(cache_dir: Optional[str] = None) -> Path:
"""
if cache_dir is None:
default_cache_dir = os.path.join(tempfile.gettempdir(), "fastembed_cache")
cache_dir = Path(os.getenv("FASTEMBED_CACHE_PATH", default_cache_dir))
cache_path = Path(os.getenv("FASTEMBED_CACHE_PATH", default_cache_dir))
else:
cache_dir = Path(cache_dir)
cache_path = Path(cache_dir)

cache_dir.mkdir(parents=True, exist_ok=True)
cache_path.mkdir(parents=True, exist_ok=True)

return cache_dir
return cache_path
10 changes: 5 additions & 5 deletions fastembed/text/e5_onnx_embedding.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,8 +22,8 @@
"size_in_GB": 1.11,
"sources": {
"hf": "xenova/paraphrase-multilingual-mpnet-base-v2",
}
}
},
},
]


Expand Down Expand Up @@ -51,8 +51,8 @@ def _preprocess_onnx_input(self, onnx_input: Dict[str, np.ndarray]) -> Dict[str,

class E5OnnxEmbeddingWorker(OnnxTextEmbeddingWorker):
def init_embedding(
self,
model_name: str,
cache_dir: str,
self,
model_name: str,
cache_dir: str,
) -> E5OnnxEmbedding:
return E5OnnxEmbedding(model_name=model_name, cache_dir=cache_dir, threads=1)
17 changes: 13 additions & 4 deletions fastembed/text/onnx_embedding.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
import os
from multiprocessing import get_all_start_methods
from typing import List, Dict, Any, Tuple, Union, Iterable, Type
from typing import List, Dict, Any, Optional, Tuple, Union, Iterable, Type

import numpy as np
import onnxruntime as ort
Expand Down Expand Up @@ -98,6 +98,15 @@
"hf": "qdrant/all-MiniLM-L6-v2-onnx",
},
},
{
"model": "nomic-ai/nomic-embed-text-v1",
"dim": 768,
"description": "8192 context length english model",
"size_in_GB": 0.54,
"sources": {
"hf": "xenova/nomic-embed-text-v1",
NirantK marked this conversation as resolved.
Show resolved Hide resolved
},
},
# {
# "model": "sentence-transformers/all-MiniLM-L6-v2",
# "dim": 384,
Expand Down Expand Up @@ -149,8 +158,8 @@ def _get_model_description(cls, model_name: str) -> Dict[str, Any]:
def __init__(
self,
model_name: str = "BAAI/bge-small-en-v1.5",
cache_dir: str = None,
threads: int = None,
cache_dir: Optional[str] = None,
threads: Optional[int] = None,
**kwargs,
):
"""
Expand Down Expand Up @@ -193,7 +202,7 @@ def embed(
self,
documents: Union[str, Iterable[str]],
batch_size: int = 256,
parallel: int = None,
parallel: Optional[int] = None,
**kwargs,
) -> Iterable[np.ndarray]:
"""
Expand Down
14 changes: 6 additions & 8 deletions fastembed/text/text_embedding.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,24 +53,22 @@ def __init__(
):
super().__init__(model_name, cache_dir, threads, **kwargs)

self.model = None
for embedding in self.EMBEDDINGS_REGISTRY:
supported_models = embedding.list_supported_models()
if any(model_name == model["model"] for model in supported_models):
self.model = embedding(model_name, cache_dir, threads, **kwargs)
break
return

if self.model is None:
raise ValueError(
f"Model {model_name} is not supported in TextEmbedding."
"Please check the supported models using `TextEmbedding.list_supported_models()`"
)
raise ValueError(
f"Model {model_name} is not supported in TextEmbedding."
"Please check the supported models using `TextEmbedding.list_supported_models()`"
)

def embed(
self,
documents: Union[str, Iterable[str]],
batch_size: int = 256,
parallel: int = None,
parallel: Optional[int] = None,
**kwargs,
) -> Iterable[np.ndarray]:
"""
Expand Down
2 changes: 1 addition & 1 deletion fastembed/text/text_embedding_base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ def embed(
self,
documents: Union[str, Iterable[str]],
batch_size: int = 256,
parallel: int = None,
parallel: Optional[int] = None,
**kwargs,
) -> Iterable[np.ndarray]:
raise NotImplementedError()
Expand Down
1 change: 1 addition & 0 deletions tests/test_onnx_embeddings.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
),
"jinaai/jina-embeddings-v2-small-en": np.array([-0.0455, -0.0428, -0.0122, 0.0613, 0.0015]),
"jinaai/jina-embeddings-v2-base-en": np.array([-0.0332, -0.0509, 0.0287, -0.0043, -0.0077]),
"nomic-ai/nomic-embed-text-v1": np.array([0.0061, 0.0103, -0.0296, -0.0242, -0.0170]),
}


Expand Down
1 change: 1 addition & 0 deletions tests/test_text_onnx_embeddings.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
),
"jinaai/jina-embeddings-v2-small-en": np.array([-0.0455, -0.0428, -0.0122, 0.0613, 0.0015]),
"jinaai/jina-embeddings-v2-base-en": np.array([-0.0332, -0.0509, 0.0287, -0.0043, -0.0077]),
"nomic-ai/nomic-embed-text-v1": np.array([0.0061, 0.0103, -0.0296, -0.0242, -0.0170]),
}


Expand Down
Loading