Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[AIC-py] hf machine translation (text to text) model parser #753

Merged
merged 2 commits into from
Jan 5, 2024
Merged

[AIC-py] hf machine translation (text to text) model parser #753

merged 2 commits into from
Jan 5, 2024

Commits on Jan 4, 2024

  1. [RFC][AIC-py] hf summarization model parser 

    Very similar to text gen

Test:

```
...
    {
      "name": "test_hf_sum",
      "input": "HMS Duncan was a D-class destroyer ...", # [contents of https://en.wikipedia.org/wiki/HMS_Duncan_(D99)]
      "metadata": {
        "model": {
          "name": "stevhliu/my_awesome_billsum_model",
          "settings": {
            "min_length": 100,
            "max_length": 200,
            "num_beams": 1
          }
        }
      }
    },
...
}
```

```
import asyncio
from aiconfig_extension_hugging_face.local_inference.text_summarization import HuggingFaceTextSummarizationTransformer

from aiconfig import AIConfigRuntime, InferenceOptions, CallbackManager

# Load the aiconfig.

mp = HuggingFaceTextSummarizationTransformer()

AIConfigRuntime.register_model_parser(mp, "stevhliu/my_awesome_billsum_model")

config = AIConfigRuntime.load("/Users/jonathan/Projects/aiconfig/cookbooks/Getting-Started/travel.aiconfig.json")
config.callback_manager = CallbackManager([])


def print_stream(data, _accumulated_data, _index):
    print(data, end="", flush=True)


async def run():
    print("Stream")
    options = InferenceOptions(stream=True, stream_callback=print_stream)
    out = await config.run("test_hf_sum", options=options)
    print("Output:\n", out)

    print("no stream")
    options = InferenceOptions(stream=False)
    out = await config.run("test_hf_sum", options=options)
    print("Output:\n", out)


asyncio.run(run())


# OUT:

Stream
Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 512). Running this sequence through the model will result in indexing errors
/opt/homebrew/Caskroom/miniconda/base/envs/aiconfig/lib/python3.10/site-packages/transformers/generation/configuration_utils.py:430: UserWarning: `num_beams` is set to 1. However, `early_stopping` is set to `True` -- this flag is only used in beam-based generation modes. You should set `num_beams>1` or unset `early_stopping`.
  warnings.warn(
/opt/homebrew/Caskroom/miniconda/base/envs/aiconfig/lib/python3.10/site-packages/transformers/generation/configuration_utils.py:449: UserWarning: `num_beams` is set to 1. However, `length_penalty` is set to `2.0` -- this flag is only used in beam-based generation modes. You should set `num_beams>1` or unset `length_penalty`.
  warnings.warn(
<pad> escorted the 13th Destroyer Flotilla in the Mediterranean and escorited the carrier Argus to Malta during the war  The ship was recalled home to be converted into an escoter destroyer in late 1942. The vessel was repaired and given a refit at Gibraltar on 16 November, and was sold for scrap later that year. The crew of the ship escores the ship to the Middle East, and the ship is a 'disaster' of the </s>Output:
 [ExecuteResult(output_type='execute_result', execution_count=0, data="<pad> escorted the 13th Destroyer Flotilla in the Mediterranean and escorited the carrier Argus to Malta during the war  The ship was recalled home to be converted into an escoter destroyer in late 1942. The vessel was repaired and given a refit at Gibraltar on 16 November, and was sold for scrap later that year. The crew of the ship escores the ship to the Middle East, and the ship is a 'disaster' of the </s>", mime_type=None, metadata={})]
no stream
Output:
 [ExecuteResult(output_type='execute_result', execution_count=0, data="escorted the 13th Destroyer Flotilla in the Mediterranean and escorited the carrier Argus to Malta during the war . The ship was recalled home to be converted into an escoter destroyer in late 1942. The vessel was repaired and given a refit at Gibraltar on 16 November, and was sold for scrap later that year. The crew of the ship escores the ship to the Middle East, and the ship is a 'disaster' of the .", mime_type=None, metadata={})]

```
    @jonathanlastmileai
    jonathanlastmileai committed Jan 4, 2024
    Configuration menu
    Copy the full SHA
    13e590f View commit details
    Browse the repository at this point in the history

  2. [AIC-py] hf mt model parser 

    text translation sans streaming. Streaming is not trivial here

Test:

```
import asyncio
from aiconfig_extension_hugging_face.local_inference.text_translation import HuggingFaceTextTranslationTransformer

from aiconfig import AIConfigRuntime, InferenceOptions, CallbackManager

# Load the aiconfig.

mp = HuggingFaceTextTranslationTransformer()

AIConfigRuntime.register_model_parser(mp, "translation_en_to_fr")

config = AIConfigRuntime.load("/Users/jonathan/Projects/aiconfig/test_hf_transl.aiconfig.json")
config.callback_manager = CallbackManager([])


def print_stream(data, _accumulated_data, _index):
    print(data, end="", flush=True)


async def run():
    # print("Stream")
    # options = InferenceOptions(stream=True, stream_callback=print_stream)
    # out = await config.run("test_hf_trans", options=options)
    # print("Output:\n", out)

    print("no stream")
    options = InferenceOptions(stream=False)
    out = await config.run("test_hf_trans", options=options)
    print("Output:\n", out)


asyncio.run(run())

```
    @jonathanlastmileai
    jonathanlastmileai committed Jan 4, 2024
    Configuration menu
    Copy the full SHA
    9fd71be View commit details
    Browse the repository at this point in the history