v0.1.14

What's Changed

• Fix to fail to save the data to cache by @SimFG in #224

Full Changelog: 0.1.13...0.1.14

v0.1.13

🎉 Introduction to new functions of GPTCache

1. Add openai audio adapter (experimental)

cache.init(pre_embedding_func=get_file_bytes)

openai.Audio.transcribe(
    model="whisper-1",
    file=audio_file
)

2. Improve data eviction implementation

In the future, users will have greater flexibility to customize eviction methods, such as by using Redis or Memcached. Currently, the default caching library is cachetools, which provides an in-memory cache. Other libraries are not currently supported, but may be added in the future.

What's Changed

• Add openai audio adapter by @shiyu22 in #220
• Improve data eviction implementation by @SimFG in #221
• Update the version to 0.1.13 by @SimFG in #222

Full Changelog: 0.1.12...0.1.13

v0.1.12

What's Changed

• Add object storage by @junjiejiangjjj in #213
• Only when the eviction policy is LRU, the data access_time is updated by @SimFG in #214
• Add logger for gptcache by @shiyu22 in #216
• The llm request can customize topk search parameters by @SimFG in #217
• Scalar return emb in ndarray by @junjiejiangjjj in #218

🎉 Introduction to new functions of GPTCache

1. The llm request can customize topk search parameters

openai.ChatCompletion.create(
    model="gpt-3.5-turbo",
    messages=[
        {"role": "user", "content": question},
    ],
    top_k=10,
)

Full Changelog: 0.1.11...0.1.12

v0.1.11

What's Changed

• Add langchain examples by @shiyu22 in #196
• Enable Python syntax highlighting by @Pouyanpi in #199
• Add openai Completion adapter by @shiyu22 in #202
• Add openai bootcamp by @shiyu22 in #207
• Add openai.Image.create in adapter by @jaelgu in #208
• Refine scalar storage by @junjiejiangjjj in #205
• Improve the langchain example by @SimFG in #211

New Contributors

• @Pouyanpi made their first contribution in #199

🎉 Introduction to new functions of GPTCache

1. Add openai complete adapter

cache.init(pre_embedding_func=get_prompt)

response = openai.Completion.create(
                model="text-davinci-003",
                prompt=question
            )

2. Add langchain and openai bootcamp

3. Add openai image adapter (experimental)

from gptcache.adapter import openai

cache.init()
cache.set_openai_key()

prompt1 = 'a cat sitting besides a dog'
size1 = '256x256'

openai.Image.create(
    prompt=prompt1,
    size=size1,
    response_format='b64_json'
    )

4. Refine storage interface

Full Changelog: 0.1.10...0.1.11

v0.1.10

What's Changed

• Update CI timeout by @binbinlv in #190
• Add kreciprocal evaluation. by @wxywb in #191
• Improve coverage testing by @SimFG in #194
• Get_scalar_data return emb by @junjiejiangjjj in #195
• Add chatmodels for langchain by @shiyu22 in #193

Full Changelog: 0.1.9...0.1.10

v0.1.9

What's Changed

• Support to import data set by @SimFG in #182
• Refine vector interface by @junjiejiangjjj in #187
• Add a pre-process function for removing the prompt info by @SimFG in #186
• Update the version to 0.1.9 by @SimFG in #189

Full Changelog: 0.1.8...0.1.9

v0.1.8

What's Changed

• Add hnswlib by @junjiejiangjjj in #167
• Update the contributing doc and remove nop post-process function by @SimFG in #172
• Update LangChainLLMs with LLM by @shiyu22 in #170
• Import package with utils by @shiyu22 in #173
• Remove faiss-cpu version to fix the python 3.10 download error by @SimFG in #174
• Update the version to 0.1.8 by @SimFG in #176

Full Changelog: 0.1.7...0.1.8

v0.1.7

What's Changed

• Fix the cache_skip param don't take effect by @SimFG in #168

Full Changelog: 0.1.5...0.1.7

v0.1.6

What's Changed

• Add more API references by @jaelgu in #128
• Update the benchmark result by @SimFG in #129
• Update docs by @shanghaikid in #130
• Update index.rst for changes in README by @jaelgu in #131
• Add workflow lint & linckcheck by @Bennu-Li in #133
• Make the cached result more similar to openAI response by @SimFG in #134
• Update the contributing doc by @SimFG in #137
• Update the test workflow by @SimFG in #146
• Rebase dev branch - 2023.4.10 by @SimFG in #162

Full Changelog: 0.1.5...0.1.6

v0.1.5

🎉 GPTCache officially released the first version.

Introduction

GPTCache is a library for creating semantic cache to store responses from LLM queries.

What's Supported

• LLM Adapter
  • Support OpenAI ChatGPT API
  • Support langchain
• Embedding
  • Disable embedding. This will turn GPTCache into a keyword-matching cache
  • Support OpenAI embedding API
  • Support ONNX with the GPTCache/paraphrase-albert-onnx model
  • Support Hugging Face embedding API
  • Support Cohere embedding API
  • Support fastText embedding API
  • Support SentenceTransformers embedding API
• Cache Storage
  • Support SQLite
  • Support PostgreSQL
  • Support MySQL
  • Support MariaDB
  • Support SQL Server
  • Support Oracle
• Vector Store
  • Support Milvus
  • Support Zilliz Cloud
  • Support FAISS
• Similarity Evaluator
  • The distance we obtain from the Vector Store
  • A model-based similarity determined using the GPTCache/albert-duplicate-onnx model from ONNX
  • Exact matches between the input request and the requests obtained from the Vector Store
  • Distance represented by applying linalg.norm from numpy to the embeddings

Full Changelog: https://github.com/zilliztech/GPTCache/commits/0.1.5