upstream-archive/fastembed
1
0
Fork 0
mirror of https://github.com/qdrant/fastembed.git synced 2026-08-03 16:41:00 -05:00
Code Issues Packages Projects Releases Wiki Activity
408 Commits 38 Branches 59 Tags
gpu-runtime-debug
Commit Graph

408 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
NirantK
6d33b85f37 * chore(.gitignore): add missing newline at the end of the file 
* chore(.gitignore): add nbs/fast-*/* to the ignore list
 2023-08-17 12:47:36 +05:30
NirantK
a664dfffd2 * docs(README.md): add details about the lightness of FastEmbed library 
* docs(README.md): mention accuracy/recall performance of FastEmbed library compared to OpenAI Ada-002
 2023-08-17 12:47:27 +05:30
NirantK
ae86299ede * feat(01_ONNX_Port.ipynb): update input_texts with sample queries and passages 2023-08-17 12:15:56 +05:30
NirantK
a2a83a772b *chore: Cleanup imports 2023-08-17 12:08:38 +05:30
NirantK
5b213ce75b * chore(pyproject.toml): update onnx version to 1.11.0 
* feat(pyproject.toml): add notebook dependency with version >=7.0.2
 2023-08-17 10:47:55 +05:30
NirantK
7b1f1b5989 * feat(pyproject.toml): add onnx to dev dependency 2023-08-17 10:40:49 +05:30
NirantK
d507f20c1d * feat(pyproject.toml): add optimum dependency 2023-08-17 10:39:29 +05:30
Nirant
e7a002f8dd Add Accurate and Fast 2023-08-17 10:36:20 +05:30
Nirant
9fb7c0678f Update README.md 2023-08-17 10:32:29 +05:30
NirantK
1ba8f79ff8 * chore(pyproject.toml): update version to 0.0.2 and description to "Fast, State of the Art Quantized Embedding Models" 2023-08-16 18:49:45 +05:30
NirantK
ff1936545c * feat(02_Usage.ipynb): add notebook file for usage examples 2023-08-16 17:28:00 +05:30
NirantK
f7a3254ce4 * refactor(embedding.py): remove print statement for decompression message 
* fix(embedding.py): add default value for onnx_providers parameter in FlagEmbedding constructor
 2023-08-16 17:27:52 +05:30
NirantK
aaf1aacf15 poetry lock fix 2023-08-16 17:27:41 +05:30
NirantK
3f2c96a6e8 * chore(pyproject.toml): remove sentence-transformers and pre-commit dependencies 
* feat(pyproject.toml): add onnxruntime-silicon dependency
 2023-08-10 22:05:12 +05:30
NirantK
6fca49781e * feat(embedding.py): add download_file_from_gcs method with documentation 2023-08-10 22:01:41 +05:30
NirantK
4ea33a7140 * docs(embedding.py): add docstring for decompress_to_cache method 
* docs(embedding.py): add docstring for FlagEmbedding class constructor
 2023-08-10 21:49:58 +05:30
NirantK
c2927becce * chore(qdrant_client.py): remove unused qdrant_client.py file 2023-08-10 21:40:53 +05:30
NirantK
d32d45c39b * fix(embedding.py): add missing comma after max_length parameter in DefaultEmbedding constructor 2023-08-10 21:40:27 +05:30
NirantK
fa3fc08528 * refactor(embedding.py): remove unused SentenceTransformersEmbedding class 
* feat(embedding.py): add NotImplementedError to abstract method encode() in Embedding class
 2023-08-10 21:39:59 +05:30
NirantK
7eba8cb187 * feat(embedding.py): add DefaultEmbedding class as a subclass of FlagEmbedding 
* fix(embedding.py): fix indentation in DefaultEmbedding class constructor
 2023-08-10 21:39:33 +05:30
NirantK
0537e80d35 * fix(embedding.py): fix linting issues and add FlagEmbedding class for handling flag embeddings 
* feat(embedding.py): add support for flag-based embeddings using ONNX models
 2023-08-10 21:39:10 +05:30
NirantK
809c60bcc9 * refactor(embedding.py): remove unnecessary cache_dir initialization in DefaultEmbedding constructor 
* feat(embedding.py): add optional cache_dir parameter to decompress_to_cache method
 2023-08-10 19:19:41 +05:30
NirantK
c9b0811b57 * chore(.gitignore): add nbs/local_cache/*/* and local_cache/*/* to ignore list 2023-08-10 19:19:28 +05:30
NirantK
3ec012a17e * refactor(embedding.py): remove unused import of tempfile 2023-08-10 19:15:03 +05:30
NirantK
8ede18cc6e * fix(embedding.py): change cache directory path from using tempfile.gettempdir() to using a local_cache directory in the current working directory 
* refactor(embedding.py): remove unnecessary check for cache directory existence and creation, as it is already handled by the mkdir() method
 2023-08-10 19:13:26 +05:30
NirantK
1f403f9a2c * chore(embedding.py): fix indentation in ONNXProviders class 
* fix(embedding.py): change default onnx_providers value to [ONNXProviders.CPU]
* fix(embedding.py): raise PermissionError with specific error message when encountering 403 status code
 2023-08-10 19:07:34 +05:30
NirantK
0b14ed84b7 * chore(embedding.py): update GPU support comment in ONNXProviders class 2023-08-10 10:56:06 +05:30
NirantK
a917a5f202 * fix(embedding.py): change default model_name from "BAAI/bge-base-en" to "BAAI/bge-small-en" 2023-08-10 10:56:00 +05:30
NirantK
4adb430bb2 * fix(embedding.py): change batch_size default value from 32 to 256 
* chore(embedding.py): add TODO comment about normalizing embeddings after all batches are done
 2023-08-10 10:50:11 +05:30
NirantK
98e95e8af6 * reformat 2023-08-10 10:47:14 +05:30
NirantK
5b46cb81e3 * refactor(embedding.py): Remove unused imports and fix typing import 
* fix(qdrant_mixin.py): Fix syntax error in dict() method call
* chore(speculative.py): Remove speculative.py file
 2023-08-10 10:46:13 +05:30
NirantK
f01a778638 * refactor(embedding.py): Remove GeneralTextEmbedding class from embedding.py 
* feat(speculative.py): Add GeneralTextEmbedding class to speculative.py
 2023-08-10 10:45:34 +05:30
NirantK
11734b2e8b * feat(embedding.py): add docstring to the encode method explaining its functionality 2023-08-10 10:44:50 +05:30
NirantK
8e876dd330 * fix(embedding.py): fix model_name format assertion error 
* feat(embedding.py): add support for custom model names and download from GCS using the model name
* refactor(embedding.py): refactor tokenizer and model initialization to use max_length and onnx_providers variables
 2023-08-10 10:44:19 +05:30
NirantK
885a738751 * feat(embedding.py): add support for custom model name, onnx providers, and max length in DefaultEmbedding constructor 2023-08-10 10:43:50 +05:30
NirantK
6cce4e0b9a * chore(01_ONNX_Port.ipynb): clean up notebook formatting and imports 
* feat(01_ONNX_Port.ipynb): add imports for pipeline and ONNX-related modules
 2023-08-10 10:43:18 +05:30
NirantK
17d05087ef * chore(usage.py): remove unused usage.py file 2023-08-10 10:42:37 +05:30
NirantK
63581a0a56 * chore(dev-requirements.txt): add google-cloud-storage to dev requirements 2023-08-10 10:42:30 +05:30
NirantK
4bfd5a13e1 * chore(embedding.py): add ONNXProviders class with CPU, GPU, and Metal execution providers 
* feat(embedding.py): add DefaultEmbedding class with model_name parameter
 2023-08-10 10:42:21 +05:30
NirantK
bc8e160657 * chore(dev-requirements.txt): add dev dependencies for jupyter, jupyter_nbextensions_configurator, black[jupyter], isort, and ruff 2023-08-05 16:40:54 +05:30
NirantK
0721f6a447 * chore(embedding.py): set default logger severity to ERROR level 2023-08-03 13:25:57 +05:30
NirantK
b353a1d7e1 * fix(qdrant_client.py): change import statement from QdrantClientMixin to QdrantAPIExtensions 
* fix(qdrant_mixin.py): change class name from QdrantClientMixin to QdrantAPIExtensions
* fix(qdrant_mixin.py): change method name
 2023-08-03 13:03:48 +05:30
NirantK
a0a57c3387 * chore: reformat with black 2023-08-03 13:02:33 +05:30
NirantK
63979dedb7 * chore(pyproject.toml): update project version to 0.0.1 and description to "Fast Embedding Creation and Simpler API for Qdrant" 
* docs(pyproject.toml): update homepage and repository URLs
* feat(pyproject.toml): add "sentence-transformers"
 2023-07-31 17:37:57 +05:30
NirantK
0a861ba5e1 * chore(pyproject.toml): update version from 0.0.1.alpha2 to 0.0.1.alpha3 2023-07-31 16:48:23 +05:30
NirantK
1c684b1208 * chore(.gitignore): add *.tar.gz to the list of ignored files 2023-07-31 16:42:45 +05:30
NirantK
6c0be8b9d1 * fix(embedding.py): fix method signature of download_file_from_gcs to include self parameter 
* fix(embedding.py): fix method signature of decompress_to_cache to include self parameter
* fix(embedding.py): update path handling in DefaultEmbedding's __init__ method
 2023-07-31 16:42:18 +05:30
NirantK
538c63d2ee * chore(qdrant_mixin.py): refactor QdrantClientMixin to use DefaultEmbedding instead of SentenceTransformersEmbedding 2023-07-31 16:28:32 +05:30
NirantK
81956c371b * chore(embedding.py): add normalize function for division by zero 
* fix(embedding.py): remove print statement
 2023-07-31 16:28:26 +05:30
NirantK
36e276fa8b * feat(embedding.py): add support for tokenization and encoding of documents using the model 
* chore(pyproject.toml): add tokenizers as a dependency
 2023-07-31 16:27:39 +05:30