dspy
dspy copied to clipboard
DSPy: The framework for programming—not prompting—language models
I believe I read somewhere that contingent on interest, there are plans for more extensive lcel integration. Please consider this interest:) Steaming is critical to our use case and other...
Is it possible to use json objects for context and answers? Is the best solution to do something like: ```python answer = dspy.OutputField(desc=""" Use the following format: {"summary": string, "commands:...
IPython is not included as a dependency, but is required to import some files: I.e ```python from dspy.teleprompt import ( BootstrapFewShotWithRandomSearch, BootstrapFewShot, LabeledFewShot, ) ``` ``` ... File "/dspy/evaluate/evaluate.py", line...
It seems that retrieval from Qdrant is limited to the following models: ```python SUPPORTED_EMBEDDING_MODELS: Dict[str, Tuple[int, models.Distance]] = { "BAAI/bge-base-en": (768, models.Distance.COSINE), "sentence-transformers/all-MiniLM-L6-v2": (384, models.Distance.COSINE), "BAAI/bge-small-en": (384, models.Distance.COSINE), "BAAI/bge-small-en-v1.5": (384,...
https://platform.openai.com/docs/assistants/overview OpenAI put Assistant API into Beta
A [recent paper](https://www.nature.com/articles/s41586-023-06924-6) by DeepMind demonstrated the use of FunSearch, an "evolutionary procedure" pairing an LLM with a systematic evaluator, on problems in extremal combinatorics (cap set) and algorithm design...
This changes allows running some large models on CPU locally. Let me know how I can improve it
Since the introduction of [PEP 621](https://peps.python.org/pep-0621), using `pyproject.toml` is [the standard way](https://packaging.python.org/en/latest/specifications/declaring-project-metadata/#declaring-project-metadata) for declaring project metadata. There seems to be little, if any, reason to keep using `setup.py` method. More...
DSPy is using `openai
Added MongoDB Atlas Retrieval Model. Currently only supports OpenAI embedding provider.