tfjs-uv

Save all deprecated TensorFlow.js environments.

uv sync
source .venv/bin/activate

tensorflowjs_converter -h

usage: TensorFlow.js model converters.
[-h]
[--input_format {keras_keras,keras,keras_saved_model,tf_hub,
tf_frozen_model,tf_saved_model,tfjs_layers_model}]
[--output_format {tfjs_graph_model,keras_keras,keras,keras_saved_model,tfjs_layers_model}]
[--signature_name SIGNATURE_NAME]
[--saved_model_tags SAVED_MODEL_TAGS]
[--quantize_float16 [QUANTIZE_FLOAT16]]
[--quantize_uint8 [QUANTIZE_UINT8]]
[--quantize_uint16 [QUANTIZE_UINT16]]
[--quantization_bytes {1,2}]
[--split_weights_by_layer] [--version]
[--skip_op_check]
[--strip_debug_ops STRIP_DEBUG_OPS]
[--use_structured_outputs_names USE_STRUCTURED_OUTPUTS_NAMES]
[--weight_shard_size_bytes WEIGHT_SHARD_SIZE_BYTES]
[--output_node_names OUTPUT_NODE_NAMES]
[--control_flow_v2 CONTROL_FLOW_V2]
[--experiments EXPERIMENTS]
[--metadata METADATA]
[input_path] [output_path]

positional arguments:
  input_path
    Path to the input file or directory. For input format "keras",
    an HDF5 (.h5) file is expected. For input format "tensorflow",
    a SavedModel directory, frozen model file, or TF-Hub module is
    expected.
  output_path
    Path for all output artifacts.

options:
  -h, --help
    show this help message and exit
  --input_format {
      keras_keras,keras,keras_saved_model,tf_hub,tf_frozen_model,
      tf_saved_model,tfjs_layers_model
    }
    Input format. For "keras", the input path can be one of the
    two following formats: - A topology+weights combined HDF5
    (e.g., generated with `tf_keras.model.save_model()` method). -
    A weights-only HDF5 (e.g., generated with Keras Model's
    `save_weights()` method). For "keras_saved_model", the
    input_path must point to a subfolder under the saved model
    folder that is passed as the argument to
    tf.contrib.save_model.save_keras_model(). The subfolder is
    generated automatically by tensorflow when saving keras model
    in the SavedModel format. It is usually named as a Unix epoch
    time (e.g., 1542212752). For "tf" formats, a SavedModel,
    frozen model, or TF-Hub module is expected.
  --output_format {tfjs_graph_model,keras_keras,keras,keras_saved_model,tfjs_layers_model}
    Output format. Default: tfjs_graph_model.
  --signature_name SIGNATURE_NAME
    Signature of the SavedModel Graph or TF-Hub module to load.
    Applicable only if input format is "tf_hub" or
    "tf_saved_model".
  --saved_model_tags SAVED_MODEL_TAGS
    Tags of the MetaGraphDef to load, in comma separated string
    format. Defaults to "serve". Applicable only if input format
    is "tf_saved_model".
  --quantize_float16 [QUANTIZE_FLOAT16]
    Comma separated list of node names to apply float16
    quantization. You can also use wildcard symbol (*) to apply
    quantization to multiple nodes (e.g., conv/*/weights). When
    the flag is provided without any nodes the default behavior
    will match all nodes.
  --quantize_uint8 [QUANTIZE_UINT8]
    Comma separated list of node names to apply 1-byte affine
    quantization. You can also use wildcard symbol (*) to apply
    quantization to multiple nodes (e.g., conv/*/weights). When
    the flag is provided without any nodes the default behavior
    will match all nodes.
  --quantize_uint16 [QUANTIZE_UINT16]
    Comma separated list of node names to apply 2-byte affine
    quantization. You can also use wildcard symbol (*) to apply
    quantization to multiple nodes (e.g., conv/*/weights). When
    the flag is provided without any nodes the default behavior
    will match all nodes.
  --quantization_bytes {1,2}
    (Deprecated) How many bytes to optionally quantize/compress
    the weights to. 1- and 2-byte quantizaton is supported. The
    default (unquantized) size is 4 bytes.
  --split_weights_by_layer
    Applicable to keras input_format only: Whether the weights
    from different layers are to be stored in separate weight
    groups, corresponding to separate binary weight files.
    Default: False.
  --version, -v
    Show versions of tensorflowjs and its dependencies
  --skip_op_check
    Skip op validation for TensorFlow model conversion.
  --strip_debug_ops STRIP_DEBUG_OPS
    Strip debug ops (Print, Assert, CheckNumerics) from graph.
  --use_structured_outputs_names USE_STRUCTURED_OUTPUTS_NAMES
    TFJS graph outputs become a tensor map with the same structure
    as the TF graph structured_outputs (only supported for
    structured_outputs of the form {key1: tensor1, key2:
    tensor2...})
  --weight_shard_size_bytes WEIGHT_SHARD_SIZE_BYTES
    Shard size (in bytes) of the weight files. Currently
    applicable only when output_format is tfjs_layers_model or
    tfjs_graph_model.
  --output_node_names OUTPUT_NODE_NAMES
    The names of the output nodes, separated by commas. E.g.,
    "logits,activations". Applicable only if input format is
    "tf_frozen_model".
  --control_flow_v2 CONTROL_FLOW_V2
    Enable control flow v2 ops, this would improve inference
    performance on models with branches or loops.
  --experiments EXPERIMENTS
    Enable experimental features, you should only enable this flag
    when using Python3 and TensorFlow nightly build.
  --metadata METADATA
    Attach user defined metadata in format key:path/metadata.json
    Separate multiple metadata files by comma.

tree saved_model

saved_model
├── assets
├── fingerprint.pb
├── xxxx.tflite
├── saved_model.pb
└── variables
    ├── variables.data-00000-of-00001
    └── variables.index

tensorflowjs_converter \
--input_format tf_saved_model \
--output_format tfjs_graph_model \
saved_model \
tfjs_model

tree tfjs_model

tfjs_model
├── group1-shard1of1.bin
└── model.json