Xi Yan
027ee2335c
delete old tests
2024-11-07 18:06:21 -08:00
Xi Yan
94a56cc3f3
register task required
2024-11-07 16:41:23 -08:00
Xi Yan
fd581c3d88
only keep 1 run_eval
2024-11-07 16:17:49 -08:00
Xi Yan
f05db9a25c
add eval_id for jobs
2024-11-07 14:30:46 -08:00
Xi Yan
51c20f9c29
api refactor
2024-11-07 13:54:26 -08:00
Xi Yan
97dcd5704c
Merge branch 'main' into eval_task_register
2024-11-07 13:08:58 -08:00
Ashwin Bharambe
694c142b89
Add provider deprecation support; change directory structure ( #397 )
...
* Add provider deprecation support; change directory structure
* fix a couple dangling imports
* move the meta_reference safety dir also
2024-11-07 13:04:53 -08:00
Xi Yan
283b5c1def
Merge branch 'main' into eval_task_register
2024-11-06 21:50:09 -08:00
Xi Yan
3f1ac29d57
test eval works
2024-11-06 21:40:38 -08:00
Xi Yan
413a1b6d8b
fix eval
2024-11-06 21:10:54 -08:00
Ashwin Bharambe
489f74a70b
Allow simpler initialization of RemoteProviderConfig; fix issue in httpx client
2024-11-06 19:19:26 -08:00
Xi Yan
56239fce90
scoring fix
2024-11-06 18:07:16 -08:00
Ashwin Bharambe
064d2a5287
Remove the safety adapter for Together; we can just use "meta-reference" ( #387 )
2024-11-06 17:36:57 -08:00
Xi Yan
c5cf9c30be
score batch
2024-11-06 17:30:46 -08:00
Xi Yan
0bce74402f
scoring test pass
2024-11-06 17:27:55 -08:00
Xi Yan
0351072531
fix scoring register
2024-11-06 17:18:16 -08:00
Xi Yan
def6d5d8ad
scoring resolve
2024-11-06 17:04:25 -08:00
Xi Yan
c53733d1a3
fixture
2024-11-06 16:41:17 -08:00
Xi Yan
00869799a1
Merge branch 'main' into eval_task_register
2024-11-06 16:34:22 -08:00
Ashwin Bharambe
7c340f0236
rename test_inference -> test_text_inference
2024-11-06 16:12:50 -08:00
Xi Yan
10eda0af59
delete unused
2024-11-06 16:08:04 -08:00
Ashwin Bharambe
3b54ce3499
remote::vllm now works with vision models
2024-11-06 16:07:17 -08:00
Xi Yan
1fe4099bd0
datasetio test
2024-11-06 16:00:38 -08:00
Xi Yan
1b7e19d5d0
Merge branch 'main' into eval_task_register
2024-11-06 15:05:46 -08:00
Ashwin Bharambe
994732e2e0
impls -> inline, adapters -> remote (#381 )
2024-11-06 14:54:05 -08:00
Ashwin Bharambe
b10e9f46bb
Enable remote::vllm ( #384 )
...
* Enable remote::vllm
* Kill the giant list of hard coded models
2024-11-06 14:42:44 -08:00
Dinesh Yeduguru
6ebd553da5
fix routing tables look up key for memory bank ( #383 )
...
Co-authored-by: Dinesh Yeduguru <dineshyv@fb.com>
2024-11-06 13:32:46 -08:00
Xi Yan
683a370d23
wip tests
2024-11-06 10:03:49 -08:00
Ashwin Bharambe
cde9bc1388
Enable vision models for (Together, Fireworks, Meta-Reference, Ollama) ( #376 )
...
* Enable vision models for Together and Fireworks
* Works with ollama 0.4.0 pre-release with the vision model
* localize media for meta_reference inference
* Fix
2024-11-05 16:22:33 -08:00
Ashwin Bharambe
7cf4c905f3
add support for remote providers in tests
2024-11-04 20:30:46 -08:00
Ashwin Bharambe
ffedb81c11
Significantly simpler and malleable test setup ( #360 )
...
* Significantly simpler and malleable test setup
* convert memory tests
* refactor fixtures and add support for composable fixtures
* Fix memory to use the newer fixture organization
* Get agents tests working
* Safety tests work
* yet another refactor to make this more general
now it accepts --inference-model, --safety-model options also
* get multiple providers working for meta-reference (for inference + safety)
* Add README.md
---------
Co-authored-by: Ashwin Bharambe <ashwin@meta.com>
2024-11-04 17:36:43 -08:00
Ashwin Bharambe
37b330b4ef
add dynamic clients for all APIs ( #348 )
...
* add dynamic clients for all APIs
* fix openapi generator
* inference + memory + agents tests now pass with "remote" providers
* Add docstring which fixes openapi generator :/
2024-10-31 14:46:25 -07:00
Ashwin Bharambe
eccd7dc4a9
Avoid warnings from pydantic for overriding schema
...
Also fix structured output in completions
2024-10-28 21:39:48 -07:00
Xi Yan
ed833bb758
[Evals API][7/n] braintrust scoring provider ( #333 )
...
* wip scoring refactor
* llm as judge, move folders
* test full generation + eval
* extract score regex to llm context
* remove prints, cleanup braintrust in this branch
* braintrust skeleton
* datasetio test fix
* braintrust provider
* remove prints
* dependencies
* change json -> class
* json -> class
* remove initialize
* address nits
* check identifier prefix
* braintrust scoring identifier check, rebase
* udpate MANIFEST
* manifest
* remove braintrust scoring_fn
* remove comments
* tests
* imports fix
2024-10-28 18:59:35 -07:00
Xi Yan
7b8748c53e
[Evals API][6/n] meta-reference llm as judge, registration for ScoringFnDefs ( #330 )
...
* wip scoring refactor
* llm as judge, move folders
* test full generation + eval
* extract score regex to llm context
* remove prints, cleanup braintrust in this branch
* change json -> class
* remove initialize
* address nits
* check identifier prefix
* udpate MANIFEST
2024-10-28 14:08:42 -07:00
Ashwin Bharambe
b7d2b83d55
Allow passing provider_registry to resolve_impls()
2024-10-28 11:58:16 -07:00
Dinesh Yeduguru
9b85d9a841
completion() for fireworks ( #329 )
2024-10-25 16:12:10 -07:00
Dinesh Yeduguru
7ec79f3b9d
completion() for together ( #324 )
...
* completion() for together
* test fixes
* fix client building
2024-10-25 14:21:12 -07:00
Xi Yan
abdf7cddf3
[Evals API][4/n] evals with generation meta-reference impl ( #303 )
...
* wip
* dataset validation
* test_scoring
* cleanup
* clean up test
* comments
* error checking
* dataset client
* test client:
* datasetio client
* clean up
* basic scoring function works
* scorer wip
* equality scorer
* score batch impl
* score batch
* update scoring test
* refactor
* validate scorer input
* address comments
* evals with generation
* add all rows scores to ScoringResult
* minor typing
* bugfix
* scoring function def rename
* rebase name
* refactor
* address comments
* Update iOS inference instructions for new quantization
* Small updates to quantization config
* Fix score threshold in faiss
* Bump version to 0.0.45
* Handle both ipv6 and ipv4 interfaces together
* update manifest for build templates
* Update getting_started.md
* chatcompletion & completion input type validation
* inclusion->subsetof
* error checking
* scoring_function -> scoring_fn rename, scorer -> scoring_fn rename
* address comments
* [Evals API][5/n] fixes to generate openapi spec (#323 )
* generate openapi
* typing comment, dataset -> dataset_id
* remove custom type
* sample eval run.yaml
---------
Co-authored-by: Dalton Flanagan <6599399+dltn@users.noreply.github.com>
Co-authored-by: Ashwin Bharambe <ashwin.bharambe@gmail.com>
2024-10-25 13:12:39 -07:00
Dinesh Yeduguru
3e1c3fdb3f
completion() for tgi ( #295 )
2024-10-24 16:02:41 -07:00
Xi Yan
cb84034567
[Evals API][3/n] scoring_functions / scoring meta-reference implementations ( #296 )
...
* wip
* dataset validation
* test_scoring
* cleanup
* clean up test
* comments
* error checking
* dataset client
* test client:
* datasetio client
* clean up
* basic scoring function works
* scorer wip
* equality scorer
* score batch impl
* score batch
* update scoring test
* refactor
* validate scorer input
* address comments
* add all rows scores to ScoringResult
* bugfix
* scoring function def rename
2024-10-24 14:52:30 -07:00
Ashwin Bharambe
ffb561070d
Support structured output for Together ( #289 )
2024-10-22 22:36:38 -07:00
Sarthak Deshpande
2e5e46d896
Added tests for persistence ( #274 )
2024-10-22 19:41:46 -07:00
Xi Yan
821810657f
[Evals API][2/n] datasets / datasetio meta-reference implementation ( #288 )
...
* skeleton dataset / datasetio
* dataset datasetio
* config
* address comments
* delete dataset_utils
* address comments
* naming fix
2024-10-22 16:12:16 -07:00
Ashwin Bharambe
c06718fbd5
Add support for Structured Output / Guided decoding ( #281 )
...
Added support for structured output in the API and added a reference implementation for meta-reference.
A few notes:
* Two formats are specified in the API: Json schema and EBNF based grammar
* Implementation only supports Json for now
We use lm-format-enhancer to provide the implementation right now but may change this especially because BNF grammars aren't supported by that library.
Fireworks has support for structured output and Together has limited supported for it too. Subsequent PRs will add these changes. We would like all our inference providers to provide structured output for llama models since it is an extremely important and highly sought-after need by the developers.
2024-10-22 12:53:34 -07:00
Anush
4c3d33e6f4
feat: Qdrant Vector index support ( #221 )
...
This PR adds support for Qdrant - https://qdrant.tech/ to be used as a vector memory.
I've unit-tested the methods to confirm that they work as intended.
To run Qdrant
```
docker run -p 6333:6333 qdrant/qdrant
```
2024-10-22 12:50:19 -07:00
Dinesh Yeduguru
1d241bf3fe
add completion() for ollama ( #280 )
2024-10-21 22:26:33 -07:00
Ashwin Bharambe
2089427d60
Make all methods async def again; add completion() for meta-reference ( #270 )
...
PR #201 had made several changes while trying to fix issues with getting the stream=False branches of inference and agents API working. As part of this, it made a change which was slightly gratuitous. Namely, making chat_completion() and brethren "def" instead of "async def".
The rationale was that this allowed the user (within llama-stack) of this to use it as:
```
async for chunk in api.chat_completion(params)
```
However, it causes unnecessary confusion for several folks. Given that clients (e.g., llama-stack-apps) anyway use the SDK methods (which are completely isolated) this choice was not ideal. Let's revert back so the call now looks like:
```
async for chunk in await api.chat_completion(params)
```
Bonus: Added a completion() implementation for the meta-reference provider. Technically should have been another PR :)
2024-10-18 20:50:59 -07:00
Ashwin Bharambe
95a96afe34
Small rename
2024-10-18 14:41:38 -07:00
Xi Yan
be3c5c034d
[bugfix] fix case for agent when memory bank registered without specifying provider_id ( #264 )
...
* fix case where memory bank is registered without provider_id
* memory test
* agents unit test
2024-10-17 17:28:17 -07:00
