Upload README.md with huggingface_hub

README.md

@@ -1,70 +1,45 @@
 ---
-library_name: transformers
+base_model: FacebookAI/xlm-mlm-100-1280
+language:
+- crk
+license: cc-by-4.0
+pipeline_tag: text-classification
 tags:
-- generated_from_trainer
-model-index:
-- name: xlm-mlm-alpha-0p15-plains-cree-en-calibrated
-  results: []
+- figurative-language
+- text-classification
+- plains-cree
+- clkd
+- low-resource
+- calibrated
 ---
 
-<!-- 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. -->
-
 # xlm-mlm-alpha-0p15-plains-cree-en-calibrated
 
-This model was trained from scratch on an unknown dataset.
-It achieves the following results on the evaluation set:
-- Loss: 5.4477
-- Macro F1: 0.4514
-- Literal F1: 0.8056
-- Idiom F1: 0.25
-- Metaphor F1: 0.0
-- Simile F1: 0.75
-
-## Model description
+FacebookAI/xlm-mlm-100-1280 calibrated on DeepSeek-annotated Bloomfield Plains Cree sentences for figurative language detection.
 
-More information needed
+## Training
 
-## Intended uses & limitations
+Low-LR calibration of a CLKD checkpoint on gold/silver Bloomfield validation data (all DeepSeek-annotated sentences).
 
-More information needed
+**Base checkpoint:** `data/clkd_xlm-mlm-alpha-0p15`
+**Literal : figurative ratio:** 3:1 | **Epochs:** 15 | **Learning rate:** 5e-06
+**Hardware:** 1× NVIDIA A100 40 GB
 
-## Training and evaluation data
+**Labels:** `literal` (0), `idiom` (1), `metaphor` (2), `simile` (3)
 
-More information needed
+## Intended use
 
-## Training procedure
+Final deployed checkpoint for figurative language detection in Plains Cree.
 
-### Training hyperparameters
 
-The following hyperparameters were used during training:
-- learning_rate: 5e-06
-- train_batch_size: 8
-- eval_batch_size: 16
-- seed: 42
-- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 0.1
-- num_epochs: 15
+## Citation
 
-### Training results
+If you use this model, please cite the associated thesis/paper (TBD).
 
-| Training Loss | Epoch | Step | Validation Loss | Macro F1 | Literal F1 | Idiom F1 | Metaphor F1 | Simile F1 |
-|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:--------:|:-----------:|:---------:|
-| 1.8099        | 1.0   | 24   | 1.8060          | 0.2194   | 0.6957     | 0.1818   | 0.0         | 0.0       |
-| 1.0587        | 2.0   | 48   | 1.5870          | 0.4166   | 0.7164     | 0.2      | 0.0         | 0.75      |
-| 0.5081        | 3.0   | 72   | 1.7403          | 0.4481   | 0.7353     | 0.2      | 0.0         | 0.8571    |
-| 0.4727        | 4.0   | 96   | 2.6182          | 0.4615   | 0.7887     | 0.2      | 0.0         | 0.8571    |
-| 0.1802        | 5.0   | 120  | 3.6535          | 0.4569   | 0.7887     | 0.1818   | 0.0         | 0.8571    |
-| 0.2120        | 6.0   | 144  | 4.2568          | 0.4712   | 0.8056     | 0.2222   | 0.0         | 0.8571    |
-| 0.0492        | 7.0   | 168  | 4.6632          | 0.4514   | 0.8056     | 0.25     | 0.0         | 0.75      |
-| 0.0244        | 8.0   | 192  | 5.1437          | 0.4514   | 0.8056     | 0.25     | 0.0         | 0.75      |
-| 0.0005        | 9.0   | 216  | 5.4477          | 0.4514   | 0.8056     | 0.25     | 0.0         | 0.75      |
 
+## Data
 
-### Framework versions
+Training data includes:
+- [Bloomfield (1934) *Plains Cree Texts*](https://bloomfield.kiyanaw.net) (scraped and sentence-aligned)
+- [EdTeKLA Indigenous Languages Corpora](https://github.com/EdTeKLA/IndigenousLanguages_Corpora)
 
-- Transformers 5.12.1
-- Pytorch 2.12.1+cu130
-- Datasets 5.0.0
-- Tokenizers 0.22.2