Andre Wisplinghoff
Andre Wisplinghoff
Any updates on this one? Does anyone have a well working OCR model that can be used as a basis for manual fine-tuning with custom training data? This approach worked...
Yes, the legacy engine (`--oem 0`) gets this one right. `tesseract4 --psm 7 --oem 0 2s.png 2s-out-oem0-psm7.txt` [2s-out-oem0-psm7.txt](https://github.com/tesseract-ocr/tesseract/files/795898/2s-out-oem0-psm7.txt)
Probably not related to the actual problem, but I do not think that you can pass `baseURL`, `deploymentId` and `resourceName` as query parameters: https://weaviate.io/developers/weaviate/modules/retriever-vectorizer-modules/text2vec-openai#query-time-parameters These settings related to Azure OpenAI...
We ran into the same issue, using Weaviate 1.25.8.
It is quite easy to run your own GraphiQL application for accessing Weaviate. Not sure if it has all the functionality offered in the cloud-hosted console, but AFAIK it is...
Yeah got the same and fixed it by upgrading to latest Pydantic.