DoruC/Grounded-Segment-Anything

1# coding=utf-82# Copyright 2018 Salesforce and HuggingFace Inc. team.3# Copyright (c) 2018, NVIDIA CORPORATION.  All rights reserved.4# Licensed under the Apache License, Version 2.0 (the "License");5# you may not use this file except in compliance with the License.6# You may obtain a copy of the License at7#8#     http://www.apache.org/licenses/LICENSE-2.09#10# Unless required by applicable law or agreed to in writing, software11# distributed under the License is distributed on an "AS IS" BASIS,12# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.13# See the License for the specific language governing permissions and14# limitations under the License.15""" Salesforce CTRL configuration"""16 17from ...configuration_utils import PretrainedConfig18from ...utils import logging19 20 21logger = logging.get_logger(__name__)22 23CTRL_PRETRAINED_CONFIG_ARCHIVE_MAP = {24    "Salesforce/ctrl": "https://huggingface.co/Salesforce/ctrl/resolve/main/config.json"25}26 27 28class CTRLConfig(PretrainedConfig):29    """30    This is the configuration class to store the configuration of a [`CTRLModel`] or a [`TFCTRLModel`]. It is used to31    instantiate a CTRL model according to the specified arguments, defining the model architecture. Instantiating a32    configuration with the defaults will yield a similar configuration to that of the33    [Salesforce/ctrl](https://huggingface.co/Salesforce/ctrl) architecture from SalesForce.34 35    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the36    documentation from [`PretrainedConfig`] for more information.37 38    Args:39        vocab_size (`int`, *optional*, defaults to 246534):40            Vocabulary size of the CTRL model. Defines the number of different tokens that can be represented by the41            `inputs_ids` passed when calling [`CTRLModel`] or [`TFCTRLModel`].42        n_positions (`int`, *optional*, defaults to 256):43            The maximum sequence length that this model might ever be used with. Typically set this to something large44            just in case (e.g., 512 or 1024 or 2048).45        n_embd (`int`, *optional*, defaults to 1280):46            Dimensionality of the embeddings and hidden states.47        dff (`int`, *optional*, defaults to 8192):48            Dimensionality of the inner dimension of the feed forward networks (FFN).49        n_layer (`int`, *optional*, defaults to 48):50            Number of hidden layers in the Transformer encoder.51        n_head (`int`, *optional*, defaults to 16):52            Number of attention heads for each attention layer in the Transformer encoder.53        resid_pdrop (`float`, *optional*, defaults to 0.1):54            The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.55        embd_pdrop (`int`, *optional*, defaults to 0.1):56            The dropout ratio for the embeddings.57        layer_norm_epsilon (`float`, *optional*, defaults to 1e-06):58            The epsilon to use in the layer normalization layers59        initializer_range (`float`, *optional*, defaults to 0.02):60            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.61        use_cache (`bool`, *optional*, defaults to `True`):62            Whether or not the model should return the last key/values attentions (not used by all models).63 64 65    Examples:66 67    ```python68    >>> from transformers import CTRLConfig, CTRLModel69 70    >>> # Initializing a CTRL configuration71    >>> configuration = CTRLConfig()72 73    >>> # Initializing a model (with random weights) from the configuration74    >>> model = CTRLModel(configuration)75 76    >>> # Accessing the model configuration77    >>> configuration = model.config78    ```"""79 80    model_type = "ctrl"81    keys_to_ignore_at_inference = ["past_key_values"]82    attribute_map = {83        "max_position_embeddings": "n_positions",84        "hidden_size": "n_embd",85        "num_attention_heads": "n_head",86        "num_hidden_layers": "n_layer",87    }88 89    def __init__(90        self,91        vocab_size=246534,92        n_positions=256,93        n_embd=1280,94        dff=8192,95        n_layer=48,96        n_head=16,97        resid_pdrop=0.1,98        embd_pdrop=0.1,99        layer_norm_epsilon=1e-6,100        initializer_range=0.02,101        use_cache=True,102        **kwargs,103    ):104        self.vocab_size = vocab_size105        self.n_positions = n_positions106        self.n_embd = n_embd107        self.n_layer = n_layer108        self.n_head = n_head109        self.dff = dff110        self.resid_pdrop = resid_pdrop111        self.embd_pdrop = embd_pdrop112        self.layer_norm_epsilon = layer_norm_epsilon113        self.initializer_range = initializer_range114 115        self.use_cache = use_cache116 117        super().__init__(**kwargs)118