Upload LiteWhisperForConditionalGeneration

README.md

@@ -0,0 +1,199 @@
+---
+library_name: transformers
+tags: []
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]

config.json

@@ -0,0 +1,305 @@
+{
+  "_name_or_path": "efficient-speech/lite-whisper-large-v3-fast",
+  "activation_dropout": 0.0,
+  "activation_function": "gelu",
+  "apply_spec_augment": false,
+  "architectures": [
+    "LiteWhisperForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoConfig": "configuration_lite_whisper.LiteWhisperConfig",
+    "AutoModel": "modeling_lite_whisper.LiteWhisperForConditionalGeneration"
+  },
+  "begin_suppress_tokens": null,
+  "bos_token_id": 50257,
+  "classifier_proj_size": 256,
+  "d_model": 1280,
+  "decoder_attention_heads": 20,
+  "decoder_ffn_dim": 5120,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 32,
+  "decoder_start_token_id": 50258,
+  "dropout": 0.0,
+  "encoder_attention_heads": 20,
+  "encoder_ffn_dim": 5120,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 32,
+  "eos_token_id": 50257,
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "low_rank_config": [
+    {
+      "fc1": 320,
+      "fc2": 272,
+      "k_proj": 32,
+      "out_proj": 32,
+      "q_proj": 32,
+      "v_proj": 48
+    },
+    {
+      "fc1": 192,
+      "fc2": 224,
+      "k_proj": 64,
+      "out_proj": 112,
+      "q_proj": 64,
+      "v_proj": 128
+    },
+    {
+      "fc1": 128,
+      "fc2": 112,
+      "k_proj": 48,
+      "out_proj": 160,
+      "q_proj": 48,
+      "v_proj": 160
+    },
+    {
+      "fc1": 192,
+      "fc2": 96,
+      "k_proj": 48,
+      "out_proj": 160,
+      "q_proj": 48,
+      "v_proj": 192
+    },
+    {
+      "fc1": 192,
+      "fc2": 160,
+      "k_proj": 48,
+      "out_proj": 144,
+      "q_proj": 64,
+      "v_proj": 208
+    },
+    {
+      "fc1": 192,
+      "fc2": 160,
+      "k_proj": 48,
+      "out_proj": 144,
+      "q_proj": 48,
+      "v_proj": 192
+    },
+    {
+      "fc1": 224,
+      "fc2": 224,
+      "k_proj": 64,
+      "out_proj": 208,
+      "q_proj": 80,
+      "v_proj": 304
+    },
+    {
+      "fc1": 240,
+      "fc2": 432,
+      "k_proj": 80,
+      "out_proj": 208,
+      "q_proj": 96,
+      "v_proj": 256
+    },
+    {
+      "fc1": 288,
+      "fc2": 512,
+      "k_proj": 80,
+      "out_proj": 224,
+      "q_proj": 96,
+      "v_proj": 288
+    },
+    {
+      "fc1": 288,
+      "fc2": 512,
+      "k_proj": 80,
+      "out_proj": 224,
+      "q_proj": 96,
+      "v_proj": 288
+    },
+    {
+      "fc1": 320,
+      "fc2": 528,
+      "k_proj": 48,
+      "out_proj": 240,
+      "q_proj": 64,
+      "v_proj": 288
+    },
+    {
+      "fc1": 320,
+      "fc2": 528,
+      "k_proj": 64,
+      "out_proj": 224,
+      "q_proj": 80,
+      "v_proj": 320
+    },
+    {
+      "fc1": 352,
+      "fc2": 592,
+      "k_proj": 64,
+      "out_proj": 224,
+      "q_proj": 64,
+      "v_proj": 352
+    },
+    {
+      "fc1": 400,
+      "fc2": 688,
+      "k_proj": 48,
+      "out_proj": 240,
+      "q_proj": 64,
+      "v_proj": 288
+    },
+    {
+      "fc1": 480,
+      "fc2": 656,
+      "k_proj": 64,
+      "out_proj": 256,
+      "q_proj": 80,
+      "v_proj": 448
+    },
+    {
+      "fc1": 384,
+      "fc2": 768,
+      "k_proj": 48,
+      "out_proj": 256,
+      "q_proj": 48,
+      "v_proj": 432
+    },
+    {
+      "fc1": 400,
+      "fc2": 800,
+      "k_proj": 48,
+      "out_proj": 224,
+      "q_proj": 64,
+      "v_proj": 272
+    },
+    {
+      "fc1": 464,
+      "fc2": 784,
+      "k_proj": 80,
+      "out_proj": 288,
+      "q_proj": 96,
+      "v_proj": 368
+    },
+    {
+      "fc1": 512,
+      "fc2": 848,
+      "k_proj": 64,
+      "out_proj": 256,
+      "q_proj": 80,
+      "v_proj": 304
+    },
+    {
+      "fc1": 528,
+      "fc2": 752,
+      "k_proj": 80,
+      "out_proj": 272,
+      "q_proj": 80,
+      "v_proj": 416
+    },
+    {
+      "fc1": 400,
+      "fc2": 16,
+      "k_proj": 80,
+      "out_proj": 352,
+      "q_proj": 96,
+      "v_proj": 432
+    },
+    {
+      "fc1": 672,
+      "fc2": 960,
+      "k_proj": 96,
+      "out_proj": 336,
+      "q_proj": 112,
+      "v_proj": 448
+    },
+    {
+      "fc1": 640,
+      "fc2": 976,
+      "k_proj": 112,
+      "out_proj": 448,
+      "q_proj": 144,
+      "v_proj": 544
+    },
+    {
+      "fc1": 720,
+      "k_proj": 96,
+      "out_proj": 416,
+      "q_proj": 128,
+      "v_proj": 464
+    },
+    {
+      "fc1": 752,
+      "k_proj": 128,
+      "out_proj": 448,
+      "q_proj": 176,
+      "v_proj": 560
+    },
+    {
+      "fc1": 752,
+      "k_proj": 128,
+      "out_proj": 416,
+      "q_proj": 176,
+      "v_proj": 592
+    },
+    {
+      "fc1": 848,
+      "k_proj": 144,
+      "out_proj": 496,
+      "q_proj": 192,
+      "v_proj": 624
+    },
+    {
+      "fc1": 912,
+      "fc2": 1024,
+      "k_proj": 128,
+      "out_proj": 400,
+      "q_proj": 192,
+      "v_proj": 608
+    },
+    {
+      "fc1": 944,
+      "fc2": 944,
+      "k_proj": 144,
+      "out_proj": 448,
+      "q_proj": 208,
+      "v_proj": 592
+    },
+    {
+      "fc1": 960,
+      "fc2": 976,
+      "k_proj": 176,
+      "out_proj": 496,
+      "q_proj": 272,
+      "v_proj": 560
+    },
+    {
+      "fc1": 896,
+      "fc2": 992,
+      "k_proj": 192,
+      "out_proj": 464,
+      "q_proj": 272,
+      "v_proj": 560
+    },
+    {
+      "fc1": 800,
+      "fc2": 736,
+      "k_proj": 176,
+      "out_proj": 464,
+      "q_proj": 272,
+      "v_proj": 432
+    }
+  ],
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "max_length": null,
+  "max_source_positions": 1500,
+  "max_target_positions": 448,
+  "median_filter_width": 7,
+  "model_type": "lite-whisper",
+  "num_hidden_layers": 32,
+  "num_mel_bins": 128,
+  "pad_token_id": 50256,
+  "scale_embedding": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.46.3",
+  "use_cache": true,
+  "use_weighted_layer_sum": false,
+  "vocab_size": 51866
+}

configuration_lite_whisper.py

@@ -0,0 +1,12 @@
+from transformers import WhisperConfig
+
+class LiteWhisperConfig(WhisperConfig):
+    model_type = "lite-whisper"
+
+    def __init__(
+        self, 
+        low_rank_config: list[dict[str, int]] = None,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        self.low_rank_config = low_rank_config

generation_config.json

@@ -0,0 +1,13 @@
+{
+  "_from_model_config": true,
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "decoder_start_token_id": 50258,
+  "eos_token_id": 50257,
+  "max_length": 448,
+  "pad_token_id": 50256,
+  "transformers_version": "4.46.3"
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f8cc9ce658ae80282bbfaa955b39638d5702883f24951e2e630bf7fba97183c
+size 4864811592

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30152e675fb2a99e1a7a5af859fadc21f8546a0fd04c6536a31959898869ef43
+size 265554048

modeling_lite_whisper.py

@@ -0,0 +1,79 @@
+import torch
+import torch.utils.checkpoint
+from torch import nn
+from transformers.models.whisper.configuration_whisper import WhisperConfig
+from transformers.models.whisper.modeling_whisper import (
+    WhisperEncoderLayer,
+    WhisperEncoder,
+    WhisperModel,
+    WhisperForConditionalGeneration,
+)
+
+from .configuration_lite_whisper import LiteWhisperConfig
+
+
+class LinearLowRank(nn.Module):
+    def __init__(
+        self, 
+        in_features: int, 
+        out_features: int, 
+        low_rank_features: int,
+    ):
+        super().__init__()
+
+        self.weight1 = nn.Parameter(torch.randn(in_features, low_rank_features))
+        self.weight2 = nn.Parameter(torch.randn(low_rank_features, out_features))
+        self.bias = nn.Parameter(torch.zeros(out_features))
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return (x @ self.weight1) @ self.weight2 + self.bias
+
+
+class LiteWhisperEncoderLayer(WhisperEncoderLayer):
+    def __init__(self, config: WhisperConfig, low_rank_config: dict[str, int]):
+        super().__init__(config)
+
+        if "k_proj" in low_rank_config:
+            self.self_attn.k_proj = LinearLowRank(self.embed_dim, self.embed_dim, low_rank_config["k_proj"])
+        
+        if "v_proj" in low_rank_config:
+            self.self_attn.v_proj = LinearLowRank(self.embed_dim, self.embed_dim, low_rank_config["v_proj"])
+        
+        if "q_proj" in low_rank_config:
+            self.self_attn.q_proj = LinearLowRank(self.embed_dim, self.embed_dim, low_rank_config["q_proj"])
+        
+        if "out_proj" in low_rank_config:
+            self.self_attn.out_proj = LinearLowRank(self.embed_dim, self.embed_dim, low_rank_config["out_proj"])
+
+        if "fc1" in low_rank_config:
+            self.fc1 = LinearLowRank(self.embed_dim, config.encoder_ffn_dim, low_rank_config["fc1"])
+        
+        if "fc2" in low_rank_config:
+            self.fc2 = LinearLowRank(config.encoder_ffn_dim, self.embed_dim, low_rank_config["fc2"])
+
+
+class LiteWhisperEncoder(WhisperEncoder):
+    def __init__(self, config: WhisperConfig, low_rank_config: list[dict[str, int]]):
+        super().__init__(config)
+
+        self.layers = nn.ModuleList([
+            LiteWhisperEncoderLayer(config, low_rank_config[i]) 
+            for i in range(config.encoder_layers)
+        ])
+
+
+class LiteWhisperModel(WhisperModel):
+    def __init__(self, config: WhisperConfig, low_rank_config: list[dict[str, int]]):
+        super().__init__(config)
+
+        self.encoder = LiteWhisperEncoder(config, low_rank_config)
+
+
+class LiteWhisperForConditionalGeneration(WhisperForConditionalGeneration):
+    config_class = LiteWhisperConfig
+
+    def __init__(self, config: LiteWhisperConfig):
+        low_rank_config = getattr(config, "low_rank_config", None)
+
+        super().__init__(config)
+        self.model = LiteWhisperModel(config, low_rank_config)