knox / Distroless-1.md
0 likes
0 forks
1 files
Last active
| Image | Tags | Architecture Suffixes |
|---|---|---|
| gcr.io/distroless/static-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64, arm, s390x, ppc64le |
| gcr.io/distroless/base-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64, arm, s390x, ppc64le |
| gcr.io/distroless/base-nossl-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64, arm, s390x, ppc64le |
| gcr.io/distroless/cc-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64, arm, s390x, ppc64le |
| gcr.io/distroless/python3-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64 |
| gcr.io/distroless/java-base-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64, s390x, ppc64le |
| gcr.io/distroless/java17-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64, s390x, ppc64le |
| gcr.io/distroless/java21-debian12 | latest, nonroot, debug, debug-nonroot | amd64, arm64, ppc64le |
knox / Jina-16.py
0 likes
0 forks
1 files
Last active
| 1 | # Server |
| 2 | with Deployment(uses=TokenStreamingExecutor, port=12345, protocol='grpc') as dep: |
| 3 | dep.block() |
| 4 | |
| 5 | |
| 6 | # Client |
| 7 | async def main(): |
| 8 | client = Client(port=12345, protocol='grpc', asyncio=True) |
| 9 | async for doc in client.stream_doc( |
| 10 | on='/stream', |
knox / Jina-15.py
0 likes
0 forks
1 files
Last active
| 1 | @requests(on='/stream') |
| 2 | async def task(self, doc: PromptDocument, **kwargs) -> ModelOutputDocument: |
| 3 | input = tokenizer(doc.prompt, return_tensors='pt') |
| 4 | input_len = input['input_ids'].shape[1] |
| 5 | for _ in range(doc.max_tokens): |
| 6 | output = self.model.generate(**input, max_new_tokens=1) |
| 7 | if output[0][-1] == tokenizer.eos_token_id: |
| 8 | break |
| 9 | yield ModelOutputDocument( |
| 10 | token_id=output[0][-1], |
knox / Jina-14.py
0 likes
0 forks
1 files
Last active
| 1 | from transformers import GPT2Tokenizer, GPT2LMHeadModel |
| 2 | |
| 3 | |
| 4 | class TokenStreamingExecutor(Executor): |
| 5 | def __init__(self, **kwargs): |
| 6 | super().__init__(**kwargs) |
| 7 | self.model = GPT2LMHeadModel.from_pretrained('gpt2') |
knox / Jina-13.py
0 likes
0 forks
1 files
Last active
| 1 | from docarray import BaseDoc |
| 2 | |
| 3 | |
| 4 | class PromptDocument(BaseDoc): |
| 5 | prompt: str |
| 6 | max_tokens: int |
| 7 | |
| 8 | |
| 9 | class ModelOutputDocument(BaseDoc): |
| 10 | token_id: int |
knox / Jina-12.sh
0 likes
0 forks
1 files
Last active
| 1 | jina cloud deploy jcloud-flow.yml |
knox / Jina-11.sh
0 likes
0 forks
1 files
Last active
| 1 | jina export docker-compose flow.yml docker-compose.yml |
| 2 | docker-compose up |
knox / Jina-10.sh
0 likes
0 forks
1 files
Last active
| 1 | jina export kubernetes flow.yml ./my-k8s |
| 2 | kubectl apply -R -f my-k8s |
knox / Jina-8.yaml
0 likes
0 forks
1 files
Last active
| 1 | # config.yml |
| 2 | jtype: TextToImage |
| 3 | py_modules: |
| 4 | - executor.py |
| 5 | metas: |
| 6 | name: TextToImage |
| 7 | description: Text to Image generation Executor |