Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / tensorzero/tensorzero issues and pull requests
#424 - added Zod schemas for TensorZero gateway configuration
Pull Request -
State: closed - Opened by virajmehta 25 days ago
- 1 comment
#423 - Proposal: Implement an OpenAI-compatible endpoint for the gateway
Issue -
State: closed - Opened by virajmehta 25 days ago
- 1 comment
#422 - Consider how to integrate batch inference
Issue -
State: closed - Opened by GabrielBianconi 26 days ago
- 1 comment
#420 - Consider including a UUID7 implementation in the `tensorzero` Python package
Issue -
State: closed - Opened by GabrielBianconi 27 days ago
Labels: qol
#419 - Build a Zod schema that matches the config tree
Issue -
State: closed - Opened by virajmehta 27 days ago
Labels: feature, !
#418 - Fixed issue in python client where exception handling failed due to no body in response
Pull Request -
State: closed - Opened by virajmehta 27 days ago
#417 - Sonnet is throwing errors on GCP
Issue -
State: closed - Opened by virajmehta 28 days ago
- 2 comments
#416 - Explore how to include additional debugging information in inference provider-side errors
Issue -
State: open - Opened by virajmehta 28 days ago
- 1 comment
Labels: qol
#415 - Async Python client causes exception when handling 404 in a streaming request
Issue -
State: closed - Opened by GabrielBianconi 28 days ago
Labels: bug, !
#414 - Bump Python client version
Pull Request -
State: closed - Opened by GabrielBianconi 28 days ago
#413 - Clickhouse typescript integration
Pull Request -
State: closed - Opened by virajmehta 28 days ago
#412 - Support OIDC authentication?
Issue -
State: open - Opened by cboettig 29 days ago
- 5 comments
Labels: feature
#411 - Trying to use with Ollama
Issue -
State: closed - Opened by jeremyandrews 29 days ago
- 3 comments
#410 - Warn if variant includes inference params not supported by model/providers
Issue -
State: open - Opened by GabrielBianconi 30 days ago
Labels: qol
#409 - Add `top_p`, `presence_penalty`, and `frequency_penalty` options to inference requests
Pull Request -
State: closed - Opened by virajmehta about 1 month ago
- 1 comment
#408 - Beautify charts and update notebooks
Pull Request -
State: open - Opened by eyadashrafkh about 1 month ago
#407 - Add a recipe for fine-tuning on demonstrations using Fireworks
Pull Request -
State: closed - Opened by virajmehta about 1 month ago
- 1 comment
#406 - Consider how to include billing information for observability
Issue -
State: open - Opened by GabrielBianconi about 1 month ago
Labels: feature
#405 - Update README
Pull Request -
State: closed - Opened by GabrielBianconi about 1 month ago
#404 - Migrate JSON columns to the ClickHouse JSON data type as appropriate
Issue -
State: open - Opened by virajmehta about 1 month ago
- 1 comment
Labels: qol, low-priority
#403 - Consider adding an integration with TAU-bench
Issue -
State: open - Opened by virajmehta about 1 month ago
Labels: feature
#402 - fixed python client to do streaming inference properly
Pull Request -
State: closed - Opened by virajmehta about 1 month ago
#401 - Add a version endpoint to gateway
Issue -
State: open - Opened by virajmehta about 1 month ago
Labels: feature, qol, robustness
#400 - Add "hydrated" inputs and outputs to ClickHouse ModelInference table
Pull Request -
State: closed - Opened by virajmehta about 1 month ago
#399 - Consider the correct data model for embedding ModelInferenceRequests as we move towards an embedding gateway
Issue -
State: open - Opened by virajmehta about 1 month ago
Labels: feature, qol, robustness
#398 - Improve documentation on best-of-n and mixture-of-n examples to explain what kind of prompts need to be written
Issue -
State: open - Opened by virajmehta about 1 month ago
Labels: documentation
#397 - Add a production deployment example and also CI tests that check that it works
Pull Request -
State: closed - Opened by virajmehta about 1 month ago
#396 - Add Mixture-of-N Sampling inference-time optimization (variant type)
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#395 - Implemented tagging for `/feedback` endpoint
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#394 - Implemented migrations for tagged feedback
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#393 - Implement a mixture-of-n / fuser type variant
Issue -
State: closed - Opened by virajmehta about 2 months ago
Labels: feature, !
#392 - Implement prefilling for JSON mode for Anthropic models
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#391 - Update Fireworks fine-tuning API to use the REST interface
Issue -
State: open - Opened by virajmehta about 2 months ago
Labels: qol, robustness
#389 - Consider ways to store `logprobs`
Issue -
State: open - Opened by GabrielBianconi about 2 months ago
Labels: feature
#388 - Implement configurable retries at Variant level
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#387 - Add a feature that says "retry if JSON is noncompliant with output_schema" to JSON type functions
Issue -
State: open - Opened by virajmehta about 2 months ago
Labels: feature, qol, robustness
#385 - Integrated Google AI Studio Gemini provider
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
- 1 comment
#384 - Test Ollama integration
Issue -
State: open - Opened by GabrielBianconi about 2 months ago
- 1 comment
Labels: robustness
#383 - implement Anthropic `disable_parallel_tools` flag
Issue -
State: open - Opened by virajmehta about 2 months ago
Labels: feature
#382 - Anthropic now allows multiple consecutive user / assistant messages, remove special handling for that.
Issue -
State: open - Opened by virajmehta about 2 months ago
Labels: robustness, low-priority
#381 - Update README.md
Pull Request -
State: closed - Opened by GabrielBianconi about 2 months ago
#380 - Update README.md
Pull Request -
State: closed - Opened by GabrielBianconi about 2 months ago
#379 - Think about how to best support batched inferences with e.g. anthropic and OpenAI async APIs
Issue -
State: closed - Opened by virajmehta about 2 months ago
- 1 comment
Labels: feature
#378 - Make sure prompt caching is implemented as-needed for each of the many inference providers that support it
Issue -
State: open - Opened by virajmehta about 2 months ago
Labels: feature
#377 - Make it as easy as possible to create demonstrations for every time an LLM fails JSON schema validation
Issue -
State: open - Opened by virajmehta about 2 months ago
Labels: feature, qol
#376 - Improve client
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#375 - TensorZero python classes can't be passed back into `inference` for multistep flows
Issue -
State: closed - Opened by virajmehta about 2 months ago
Labels: qol, !
#373 - Failed to run ClickHouse migrations
Issue -
State: closed - Opened by fahdmirza about 2 months ago
- 1 comment
#372 - Investigate: gateway does not 400 on added fields in `input`
Issue -
State: open - Opened by virajmehta about 2 months ago
- 1 comment
Labels: bug
#371 - Fix paths in examples
Pull Request -
State: closed - Opened by GabrielBianconi about 2 months ago
#370 - added trailing writes to feedback endpoint for latency reasons
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#369 - Clean up chess example
Pull Request -
State: closed - Opened by GabrielBianconi about 2 months ago
#368 - Fix openai json mode for non-strict
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#367 - Bump python client version
Pull Request -
State: closed - Opened by virajmehta about 2 months ago
#366 - Rename BoN references & examples to be extra clear
Pull Request -
State: closed - Opened by GabrielBianconi 2 months ago
#365 - Add `inference_params` overrides for new variant types
Issue -
State: open - Opened by GabrielBianconi 2 months ago
Labels: robustness
#364 - Dicl recipe
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#363 - Allows API keys to be passed at inference time
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#362 - Error handling within Variants and ModelProviders gives 500 when 400 is more appropriate
Issue -
State: open - Opened by virajmehta 2 months ago
Labels: robustness
#361 - Add experimental dynamic in-context learning variant type to gateway
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#360 - Enforce that DICL variants have data in the `DynamicInContextLearningExamples` table at validation time
Issue -
State: open - Opened by virajmehta 2 months ago
Labels: robustness
#359 - together has added usage to their streaming API
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#358 - Add support for multiple generations per model inference for best-of-n sampling
Issue -
State: open - Opened by GabrielBianconi 2 months ago
Labels: feature, qol
#357 - Implement configurable retries for variants
Issue -
State: closed - Opened by virajmehta 2 months ago
- 1 comment
Labels: feature, qol, robustness, !, NOW
#356 - Internal embeddings
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#355 - Changed name of dicl table to spell out the acronym
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#354 - Fix directory name
Pull Request -
State: closed - Opened by GabrielBianconi 2 months ago
#353 - `/feedback` endpoint should support batched writes
Issue -
State: open - Opened by virajmehta 2 months ago
Labels: feature, qol
#352 - Explore TensorZero automatically generating Pydantic models for the output schema of a given function
Issue -
State: open - Opened by virajmehta 2 months ago
Labels: feature, qol
#351 - Implemented migration that creates DiclExample table
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#350 - Validate IDs for feedback API calls beyond demonstrations
Issue -
State: open - Opened by GabrielBianconi 2 months ago
Labels: qol, robustness
#349 - Consider adding more data to ClickHouse so it's self-sufficient
Issue -
State: closed - Opened by GabrielBianconi 2 months ago
- 3 comments
Labels: qol, robustness, !
#348 - Allow demonstrations for dynamic tools & output schema
Issue -
State: open - Opened by virajmehta 2 months ago
Labels: feature
#347 - Adds an example of fine-tuning with demonstration data for NER
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#346 - Add `strict` parameter for OpenAI's JSON Mode
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#345 - Add `strict` parameter for OpenAI's JSON Mode
Issue -
State: closed - Opened by GabrielBianconi 2 months ago
Labels: bug, !, NOW
#344 - Add an example and/or recipe showcasing fine-tuning with demonstrations
Issue -
State: closed - Opened by GabrielBianconi 2 months ago
- 1 comment
Labels: documentation, !
#343 - Validate demonstrations
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#342 - Add a figure showing observed performance in each example
Issue -
State: open - Opened by virajmehta 2 months ago
Labels: feature
#341 - Chess example
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#340 - Bump version for 2024.09.2 release
Pull Request -
State: closed - Opened by GabrielBianconi 2 months ago
#339 - moved config into fixtures from root dir
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#338 - Best-of-N sampling doesn't correctly handle JSON mode for Anthropic models
Issue -
State: open - Opened by virajmehta 2 months ago
- 2 comments
Labels: bug
#337 - Get rid of top-level `config/`
Issue -
State: closed - Opened by GabrielBianconi 2 months ago
Labels: documentation, qol, !
#336 - made chat completion variant weights optional and added a unit test
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#335 - Error paths can be misleading for enums in config
Issue -
State: open - Opened by virajmehta 2 months ago
Labels: bug
#334 - added a deny-unknown-fields input
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#333 - Input struct needs to deny unknown fields
Issue -
State: closed - Opened by virajmehta 2 months ago
Labels: robustness, !
#332 - Implemented synchronous python client
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#331 - Update examples for Tutorial V2
Pull Request -
State: closed - Opened by GabrielBianconi 2 months ago
#330 - Implement small API changes for OpenAI o1 models
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#329 - Implement a new Variant type for best-of-n sampling
Pull Request -
State: closed - Opened by virajmehta 2 months ago
- 1 comment
#327 - Fix inference params
Pull Request -
State: closed - Opened by virajmehta 2 months ago
#326 - Update README.md
Pull Request -
State: closed - Opened by GabrielBianconi 2 months ago
#325 - Add support and testing for the O-1 series of models from OpenAI
Issue -
State: closed - Opened by virajmehta 2 months ago
Labels: !
#324 - Refactor variant into directory structure and put validation into individual configs
Pull Request -
State: closed - Opened by virajmehta 3 months ago
#323 - Update README.md
Pull Request -
State: closed - Opened by GabrielBianconi 3 months ago
#322 - bumped python client version
Pull Request -
State: closed - Opened by virajmehta 3 months ago
#321 - Update README.md
Pull Request -
State: closed - Opened by GabrielBianconi 3 months ago
#320 - Implements a ModelProvider `GCPVertexAnthropicProvider` that allows T0 to call Anthropic models via vertex
Pull Request -
State: closed - Opened by virajmehta 3 months ago