Commit 4c7838a

jetstream authors

committed

internal change

PiperOrigin-RevId: 720730187

1 parent 3fe314e commit 4c7838aCopy full SHA for 4c7838a

171 files changed

+379

-9899

lines changed

benchmarks
experimental/jax
- README.md
- inference
  - entrypoint
  - kernel
    - __init__.py
    - attention_ops.py
    - attention/tpu
    - collective_matmul_ops.py
    - linear/tpu
      - collective_matmul.py
  - model
    - __init__.py
    - llama.py
    - management
    - postprocess.py
    - sampling
      - sampler.py
  - nn
  - parallel
  - runtime
  - server
    - __init__.py
    - simple_server.py
  - utils
    - __init__.py
    - pytree_utils.py
- requirements.txt
- tests
  - kernel
    - attention/tpu
      - test_chunked_prefill_attention.py
    - linear/tpu
      - test_collective_matmul_impl.py
  - model
    - test_llama.py
  - nn
  - parallel
    - test_operations.py
google3/third_party/py/jetstream
- .github
  - CODEOWNERS
  - workflows
    - e2e_tests.yaml
    - release.yaml
    - scripts
      - create_release.js
    - unit_tests.yaml
- .gitignore
- AUTHORS
- CONTRIBUTING.md
- LICENSE
- MANIFEST.in
- Makefile
- README.md
- __init__.py
- benchmarks
  - README.md
  - __init__.py
  - benchmark_serving.py
  - eval_accuracy.py
  - open_orca_gpt4_tokenized_llama.calibration_1000.pkl
  - requirements.in
- core
  - README.md
  - __init__.py
  - config_lib.py
  - implementations
    - __init__.py
    - mock
  - metrics
    - __init__.py
    - prometheus.py
  - orchestrator.py
  - proto
  - server_lib.py
  - utils
- docs
  - observability-prometheus-metrics-in-jetstream-server.md
  - online-inference-with-maxtext-engine.md
  - profiling-with-jax-profiler-and-tensorboard.md
- engine
  - README.md
  - __init__.py
  - engine_api.py
  - mock_engine.py
  - mock_utils.py
  - sampling_utils.py
  - token_utils.py
  - tokenizer.proto
  - tokenizer_api.py
  - tokenizer_pb2.py
  - tokenizer_pb2_grpc.py
  - warmup_utils.py
- entrypoints
  - __init__.py
  - config.py
  - http
- external_tokenizers
  - __init__.py
  - llama3
    - __init__.py
    - llama3_tokenizer.py
- license_preamble.txt
- pylintrc
- requirements.txt
- setup.py
- tests
  - __init__.py
  - core
  - engine
    - __init__.py
    - external_tokenizers
      - llama2
        tokenizer.model
      - llama3
        tokenizer.model
    - test_mock_engine.py
    - test_sampling_utils.py
    - test_token_utils.py
    - test_utils.py
  - entrypoints
    - __init__.py
    - http
      - __init__.py
      - test_api_server.py
- tools
  - load_tester.py
  - maxtext
    - model_ckpt_conversion.sh
    - model_ckpt_finetune_with_aqt.sh
  - proxy_dev
    - base.Dockerfile
    - dev.Dockerfile
  - requester.py
jetstream
- core/metrics
  - prometheus.py
- external_tokenizers/llama3
  - __init__.py

171 files changed

+379

-9899

lines changed

`‎benchmarks/metrics.py‎`

This file was deleted.

Comments

(0)