Finetuned mt5_small_lg_en model

README.md

@@ -0,0 +1,90 @@
+---
+base_model: mt5_small/lg_en
+tags:
+- generated_from_trainer
+metrics:
+- bleu
+model-index:
+- name: mt5_small_lg_en
+  results: []
+---
+
+<!-- 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. -->
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/hnamuwaya-makerere-university-business-school/mt5_small_lg_en/runs/zsfbh00n)
+# mt5_small_lg_en
+
+This model is a fine-tuned version of [mt5_small/lg_en](https://huggingface.co/mt5_small/lg_en) on the Luganda Informal Data dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2071
+- Bleu: 1.1669
+- Gen Len: 6.6138
+
+## 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: 2e-05
+- train_batch_size: 16
+- eval_batch_size: 16
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- num_epochs: 30
+
+### Training results
+
+| Training Loss | Epoch | Step  | Validation Loss | Bleu   | Gen Len |
+|:-------------:|:-----:|:-----:|:---------------:|:------:|:-------:|
+| 1.2558        | 1.0   | 848   | 0.2899          | 0.0653 | 16.1851 |
+| 0.3023        | 2.0   | 1696  | 0.2764          | 0.0872 | 12.2714 |
+| 0.289         | 3.0   | 2544  | 0.2681          | 0.1524 | 9.4625  |
+| 0.2825        | 4.0   | 3392  | 0.2623          | 0.1648 | 8.42    |
+| 0.2766        | 5.0   | 4240  | 0.2564          | 0.2707 | 8.8613  |
+| 0.2695        | 6.0   | 5088  | 0.2507          | 0.3064 | 8.2628  |
+| 0.2661        | 7.0   | 5936  | 0.2454          | 0.314  | 8.3656  |
+| 0.2582        | 8.0   | 6784  | 0.2408          | 0.5769 | 8.2283  |
+| 0.2536        | 9.0   | 7632  | 0.2367          | 0.4428 | 7.6052  |
+| 0.2514        | 10.0  | 8480  | 0.2332          | 0.5161 | 6.9993  |
+| 0.248         | 11.0  | 9328  | 0.2296          | 0.6246 | 7.1652  |
+| 0.2432        | 12.0  | 10176 | 0.2268          | 0.6372 | 7.006   |
+| 0.2393        | 13.0  | 11024 | 0.2244          | 0.681  | 6.7001  |
+| 0.2367        | 14.0  | 11872 | 0.2216          | 0.7667 | 6.8613  |
+| 0.2339        | 15.0  | 12720 | 0.2193          | 0.7835 | 6.8739  |
+| 0.2313        | 16.0  | 13568 | 0.2178          | 0.7668 | 6.6861  |
+| 0.2307        | 17.0  | 14416 | 0.2160          | 0.81   | 6.7837  |
+| 0.2279        | 18.0  | 15264 | 0.2145          | 1.0551 | 6.7193  |
+| 0.2258        | 19.0  | 16112 | 0.2135          | 1.0511 | 6.6828  |
+| 0.2245        | 20.0  | 16960 | 0.2120          | 0.8869 | 6.7757  |
+| 0.2226        | 21.0  | 17808 | 0.2112          | 0.8999 | 6.6948  |
+| 0.2216        | 22.0  | 18656 | 0.2104          | 0.9144 | 6.6264  |
+| 0.222         | 23.0  | 19504 | 0.2094          | 0.9253 | 6.6317  |
+| 0.2202        | 24.0  | 20352 | 0.2090          | 0.9439 | 6.5109  |
+| 0.2199        | 25.0  | 21200 | 0.2083          | 0.9589 | 6.6549  |
+| 0.2187        | 26.0  | 22048 | 0.2079          | 0.9446 | 6.6138  |
+| 0.2186        | 27.0  | 22896 | 0.2076          | 0.9708 | 6.6065  |
+| 0.218         | 28.0  | 23744 | 0.2074          | 0.966  | 6.5707  |
+| 0.2173        | 29.0  | 24592 | 0.2072          | 1.1663 | 6.6085  |
+| 0.2181        | 30.0  | 25440 | 0.2071          | 1.1669 | 6.6138  |
+
+
+### Framework versions
+
+- Transformers 4.42.3
+- Pytorch 2.1.2
+- Datasets 2.20.0
+- Tokenizers 0.19.1

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "decoder_start_token_id": 0,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.42.3"
+}