init commit

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+checkpoints/ filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,146 @@
+---
+pipeline_tag: translation
+language:
+- multilingual
+- af
+- am
+- ar
+- as
+- az
+- be
+- bg
+- bn
+- br
+- bs
+- ca
+- cs
+- cy
+- da
+- de
+- el
+- en
+- eo
+- es
+- et
+- eu
+- fa
+- fi
+- fr
+- fy
+- ga
+- gd
+- gl
+- gu
+- ha
+- he
+- hi
+- hr
+- hu
+- hy
+- id
+- is
+- it
+- ja
+- jv
+- ka
+- kk
+- km
+- kn
+- ko
+- ku
+- ky
+- la
+- lo
+- lt
+- lv
+- mg
+- mk
+- ml
+- mn
+- mr
+- ms
+- my
+- ne
+- nl
+- 'no'
+- om
+- or
+- pa
+- pl
+- ps
+- pt
+- ro
+- ru
+- sa
+- sd
+- si
+- sk
+- sl
+- so
+- sq
+- sr
+- su
+- sv
+- sw
+- ta
+- te
+- th
+- tl
+- tr
+- ug
+- uk
+- ur
+- uz
+- vi
+- xh
+- yi
+- zh
+license: apache-2.0
+base_model:
+- FacebookAI/xlm-roberta-large
+---
+
+# COMET-instant-self-confidence
+
+This model is based on [COMET-early-exit](https://github.com/zouharvi/COMET-early-exit), which is a fork but not compatible with original Unbabel's COMET.
+To run the model, you need to first install this version of COMET either with:
+```bash
+pip install "git+https://github.com/zouharvi/COMET-early-exit#egg=comet-early-exit&subdirectory=comet_early_exit"
+```
+or in editable mode:
+```bash
+git clone https://github.com/zouharvi/COMET-early-exit.git
+cd COMET-early-exit
+pip3 install -e comet_early_exit
+```
+
+This model specifically makes prediction at each of the 25 layers, both the score and the confidence.
+This time, the confidence is the absolute error with respect to the final layer's prediction.
+```python
+model = comet_early_exit.load_from_checkpoint(comet_early_exit.download_model("zouharvi/COMET-instant-self-confidence"))
+data = [
+    {
+        "src": "Can I receive my food in 10 to 15 minutes?",
+        "mt": "Moh bych obdržet jídlo v 10 do 15 minut?",
+    },
+    {
+        "src": "Can I receive my food in 10 to 15 minutes?",
+        "mt": "Mohl bych dostat jídlo během 10 či 15 minut?",
+    }
+]
+model_output = model.predict(data, batch_size=8, gpus=1)
+
+# print predictions at 5th, 12th, and last layer
+print("scores", model_output["scores"][0][5], model_output["scores"][0][12], model_output["scores"][0][-1])
+print("estimated errors", model_output["confidences"][0][5], model_output["confidences"][0][12], model_output["confidences"][0][-1])
+
+# two top-level outputs
+assert len(model_output["scores"]) == 2 and len(model_output["confidences"]) == 2
+# each output contains prediction per each layer
+assert all(len(l) == 25 for l in model_output["scores"]) and all(len(l) == 25 for l in model_output["confidences"])
+```
+Outputs (formatted):
+```
+scores            75.60  86.60  85.74
+estimated errors  10.48   3.52   0.83
+```

checkpoints/model.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20d25859e408bba935d110b36229dd857e3eba7b6fa922aaada7b7b998e6a89c
+size 2277649338

hparams.yaml

@@ -0,0 +1,29 @@
+activations: Tanh
+batch_size: 32
+class_identifier: earlyexitconfmulti_extra
+confidence_target: last
+dropout: 0.1
+encoder_learning_rate: 1.0e-06
+encoder_model: XLM-RoBERTa
+final_activation: null
+hidden_sizes:
+- 2048
+- 1024
+keep_embeddings_frozen: true
+layer: mix
+layer_norm: false
+layer_transformation: sparsemax
+layerwise_decay: 0.95
+learning_rate: 1.5e-05
+load_pretrained_weights: true
+local_files_only: false
+loss: mse
+nr_frozen_epochs: 0.3
+optimizer: AdamW
+pool: cls
+pretrained_model: xlm-roberta-large
+train_data:
+- data/csv/train_da.csv
+validation_data:
+- data/csv/dev_da.csv
+warmup_steps: 0