This folder contains simple command-line tools for easily trying out the C++ Text Task APIs.
You will need:
- a TFLite bert based question answerer model from model maker. (e.g. mobilebert or albert available on TensorFlow Hub).
In the console, run:
# Download the model:
curl \
-L 'https://tfhub.dev/tensorflow/lite-model/mobilebert/1/metadata/1?lite-format=tflite' \
-o /tmp/mobilebert.tflite
# Run the classification tool:
bazel run -c opt \
tensorflow_lite_support/examples/task/text/desktop:bert_question_answerer_demo -- \
--model_path=/tmp/mobilebert.tflite \
--question="Where is Amazon rainforest?" \
--context="The Amazon rainforest, alternatively, the Amazon Jungle, also known in \
English as Amazonia, is a moist broadleaf tropical rainforest in the Amazon \
biome that covers most of the Amazon basin of South America. This basin \
encompasses 7,000,000 km2 (2,700,000 sq mi), of which \
5,500,000 km2 (2,100,000 sq mi) are covered by the rainforest. This region \
includes territory belonging to nine nations."In the console, you should get:
Time cost to answer the input question on CPU: 783 ms
answer[0]: 'South America.'
logit: 1.84847, start_index: 39, end_index: 40
answer[1]: 'most of the Amazon basin of South America.'
logit: 1.2921, start_index: 34, end_index: 40
answer[2]: 'the Amazon basin of South America.'
logit: -0.0959535, start_index: 36, end_index: 40
answer[3]: 'the Amazon biome that covers most of the Amazon basin of South America.'
logit: -0.498558, start_index: 28, end_index: 40
answer[4]: 'Amazon basin of South America.'
logit: -0.774266, start_index: 37, end_index: 40
You will need:
- a TFLite text classification model with certain format. (e.g. movie_review_model, a model to classify movie reviews).
In the console, run:
# Download the model:
curl \
-L 'https://storage.googleapis.com/download.tensorflow.org/models/tflite/text_classification/text_classification_v2.tflite' \
-o /tmp/movie_review.tflite
# Run the detection tool:
bazel run -c opt \
tensorflow_lite_support/examples/task/text/desktop:nl_classifier_demo -- \
--model_path=/tmp/movie_review.tflite \
--text="What a waste of my time."In the console, you should get:
Time cost to classify the input text on CPU: 0.088 ms
category[0]: 'Negative' : '0.81313'
category[1]: 'Positive' : '0.18687'
TODO(b/163086702): Update the links to models with metadata attached.
You will need:
- a Bert based TFLite text classification model from model maker. (e.g. movie_review_model available on TensorFlow Hub).
In the console, run:
# Download the model:
curl \
-L 'https://url/to/bert/nl/classifier' \
-o /tmp/bert_movie_review.tflite
# Run the segmentation tool:
bazel run -c opt \
tensorflow_lite_support/examples/task/text/desktop:bert_nl_classifier_demo -- \
--model_path=/tmp/bert_movie_review.tflite \
--text="it's a charming and often affecting journey"In the console, you should get:
Time cost to classify the input text on CPU: 491 ms
category[0]: 'negative' : '0.00006'
category[1]: 'positive' : '0.99994'
You will need:
- a universal sentence encoder QA model from TensorFlow Hub.
In the console, run:
# Download the model:
curl \
-L 'https://tfhub.dev/google/lite-model/universal-sentence-encoder-qa-ondevice/1?lite-format=tflite' \
-o /tmp/universal_sentence_encoder_qa_with_metadata.tflite
# Run UniversalSentenceEncoderQA model:
bazel run -c opt \
tensorflow_lite_support/examples/task/text/desktop:universal_sentence_encoder_qa_demo -- \
--model_path=/tmp/universal_sentence_encoder_qa_with_metadata.tfliteIn the console, you should get:
Input questions: How are you feeling today?
Output answers 0: I'm not feeling very well. Score: 14.9595
Output answers 2: He looks good. Score: 8.80944
Output answers 1: Paris is the capital of France. Score: 5.63752
You will need:
- a TFLite text embedder model such as the universal sentence encoder QA model from TensorFlow Hub.
The TextEmbedder demo tool takes two sentences as inputs, and outputs the cosine similarity between their embeddings.
In the console, run:
# Download the model:
curl \
-L 'https://tfhub.dev/google/lite-model/universal-sentence-encoder-qa-ondevice/1?lite-format=tflite' \
-o /tmp/universal_sentence_encoder_qa_with_metadata.tflite
# Run the embedder tool:
bazel run -c opt \
tensorflow_lite_support/examples/task/text/desktop:text_embedder_demo -- \
--model_path=/tmp/universal_sentence_encoder_qa_with_metadata.tflite \
--l2_normalize \
--first_sentence="It was a very sunny day." \
--second_sentence="The sun was shining brightly."In the console, you should get:
Cosine similarity: 0.952549
You will need:
- a TFLite text embedder model such as the universal sentence encoder QA model from TensorFlow Hub,
- an index built from that embedder model using Model Maker.
Model Maker also provides the ability to add the index directly to the embedder
model metadata. The demo also supports this : just omit the --index_path
argument.
In this example, we'll be using a test index built from the universal sentence encoder QA model, which only contains 5 embeddings.
In the console, run:
# Download the model:
curl \
-L 'https://tfhub.dev/google/lite-model/universal-sentence-encoder-qa-ondevice/1?lite-format=tflite' \
-o /tmp/universal_sentence_encoder_qa_with_metadata.tflite
# Run the searcher tool:
bazel run -c opt \
tensorflow_lite_support/examples/task/text/desktop:text_searcher_demo -- \
--model_path=/tmp/universal_sentence_encoder_qa_with_metadata.tflite \
--l2_normalize \
--index_path=$(pwd)/third_party/tensorflow_lite_support/cc/test/testdata/task/text/universal_sentence_encoder_index.ldb \
--input_sentence="The sun was very bright."In the console, you should get:
Results:
Rank#0:
metadata: The sun was shining on that day.
distance: 0.04618
Rank#1:
metadata: It was a sunny day.
distance: 0.10856
Rank#2:
metadata: The weather was excellent.
distance: 0.15223
Rank#3:
metadata: The cat is chasing after the mouse.
distance: 0.34271
Rank#4:
metadata: He was very happy with his newly bought car.
distance: 0.37703