Model save

README.md

@@ -1,71 +1,84 @@
 ---
-license: mit
-language:
-- en
-datasets:
-- speechbrain/LoquaciousSet
-base_model:
-- openai/whisper-large-v3-turbo
-- HuggingFaceTB/SmolLM3-3B
-pipeline_tag: automatic-speech-recognition
+library_name: transformers
 tags:
-- asr
-- speech-recognition
-- audio
-- smollm
-- whisper
-- mlp
+- generated_from_trainer
+model-index:
+- name: tiny-audio
+  results: []
 ---
 
-# Tiny Audio
-
-A speech recognition model trained in 24 hours on a single GPU for ~$12. Built with the [Tiny Audio](https://github.com/alexkroman/tiny-audio) codebase—a minimal, hackable framework for training ASR models.
-
-## Architecture
-
-```
-Audio (16kHz) → Whisper Encoder (frozen) → MLP Projector (trained) → SmolLM3-3B (frozen) → Text
-```
-
-**MLP Projector:**
-- Convolutional downsampling: 4x sequence compression via two stride-2 conv layers
-- Linear (1280 → 2048) → GELU → Linear (2048 → 2048)
-- Output normalization: RMSNorm
-
-## Training Details
-
-| | |
-|---|---|
-| **Dataset** | LoquaciousSet (25,000 hours) |
-| **Hardware** | Single NVIDIA A40 40GB |
-| **Training Time** | ~24 hours |
-| **Cost** | ~$12 |
-| **Trainable Parameters** | ~12M (projector only) |
-
-## Performance
-
-**Word Error Rate (WER): 12.14%** on LoquaciousSet test set.
-
-
-## Usage
-
-```python
-from transformers import pipeline
-
-pipe = pipeline("automatic-speech-recognition", model="mazesmazes/tiny-audio", trust_remote_code=True)
-
-result = pipe("path/to/audio.wav")
-print(result["text"])
-```
-
-## Limitations
-
-- English only
-- Optimized for 16kHz audio; other sample rates are resampled automatically
-- Performance may degrade on heavily accented speech, noisy environments, or domain-specific jargon
-- Maximum audio length limited by context window
-
-## Learn More
-
-- **[Train your own model](https://github.com/alexkroman/tiny-audio)** — The full codebase with training scripts
-- **[Free 3.5-hour course](https://github.com/alexkroman/tiny-audio/blob/main/docs/course/0-course-overview.md)** — Build your own ASR system from scratch
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# tiny-audio
+
+This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.3030
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 0.0002
+- train_batch_size: 20
+- eval_batch_size: 20
+- seed: 42
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 40
+- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 500
+- num_epochs: 1
+
+### Training results
+
+| Training Loss | Epoch  | Step  | Validation Loss |
+|:-------------:|:------:|:-----:|:---------------:|
+| 0.5056        | 0.0372 | 1000  | 0.4229          |
+| 0.4725        | 0.0745 | 2000  | 0.3813          |
+| 0.4602        | 0.1117 | 3000  | 0.3660          |
+| 0.4293        | 0.1490 | 4000  | 0.3542          |
+| 0.4113        | 0.1862 | 5000  | 0.3468          |
+| 0.4153        | 0.2234 | 6000  | 0.3404          |
+| 0.3996        | 0.2607 | 7000  | 0.3337          |
+| 0.4089        | 0.2979 | 8000  | 0.3294          |
+| 0.3743        | 0.3352 | 9000  | 0.3250          |
+| 0.4082        | 0.3724 | 10000 | 0.3214          |
+| 0.3679        | 0.4096 | 11000 | 0.3193          |
+| 0.3792        | 0.4469 | 12000 | 0.3185          |
+| 0.3723        | 0.4841 | 13000 | 0.3140          |
+| 0.3696        | 0.5214 | 14000 | 0.3109          |
+| 0.3920        | 0.5586 | 15000 | 0.3104          |
+| 0.3827        | 0.5958 | 16000 | 0.3070          |
+| 0.3593        | 0.6331 | 17000 | 0.3054          |
+| 0.3832        | 0.6703 | 18000 | 0.3056          |
+| 0.3772        | 0.7076 | 19000 | 0.3038          |
+| 0.3887        | 0.7448 | 20000 | 0.3035          |
+| 0.3666        | 0.7821 | 21000 | 0.3034          |
+| 0.3690        | 0.8193 | 22000 | 0.3033          |
+| 0.3568        | 0.8565 | 23000 | 0.3031          |
+| 0.3809        | 0.8938 | 24000 | 0.3029          |
+| 0.3736        | 0.9310 | 25000 | 0.3032          |
+| 0.3553        | 0.9683 | 26000 | 0.3030          |
+
+
+### Framework versions
+
+- Transformers 5.0.0.dev0
+- Pytorch 2.8.0+cu128
+- Datasets 3.6.0
+- Tokenizers 0.22.1