Update repository
Browse files- README.md +6 -0
- metrics.json +23 -23
- vocabulary/non_padded_namespaces.txt +1 -1
- weights.th +3 -0
README.md
CHANGED
|
@@ -3,4 +3,10 @@ tags:
|
|
| 3 |
- allennlp
|
| 4 |
---
|
| 5 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
This is an implementation of the BiDAF model with GloVe embeddings. The basic layout is pretty simple: encode words as a combination of word embeddings and a character-level encoder, pass the word representations through a bi-LSTM/GRU, use a matrix of attentions to put question information into the passage word representations (this is the only part that is at all non-standard), pass this through another few layers of bi-LSTMs/GRUs, and do a softmax over span start and span end.
|
|
|
|
| 3 |
- allennlp
|
| 4 |
---
|
| 5 |
|
| 6 |
+
# TODO: Fill this model card
|
| 7 |
+
---
|
| 8 |
+
tags:
|
| 9 |
+
- allennlp
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
This is an implementation of the BiDAF model with GloVe embeddings. The basic layout is pretty simple: encode words as a combination of word embeddings and a character-level encoder, pass the word representations through a bi-LSTM/GRU, use a matrix of attentions to put question information into the passage word representations (this is the only part that is at all non-standard), pass this through another few layers of bi-LSTMs/GRUs, and do a softmax over span start and span end.
|
metrics.json
CHANGED
|
@@ -1,27 +1,27 @@
|
|
| 1 |
{
|
| 2 |
-
"best_epoch":
|
| 3 |
-
"peak_worker_0_memory_MB":
|
| 4 |
"peak_gpu_0_memory_MB": 2483.1318359375,
|
| 5 |
-
"training_duration": "1:
|
| 6 |
-
"epoch":
|
| 7 |
-
"training_start_acc": 0.
|
| 8 |
-
"training_end_acc": 0.
|
| 9 |
-
"training_span_acc": 0.
|
| 10 |
-
"training_em": 0.
|
| 11 |
-
"training_f1": 0.
|
| 12 |
-
"training_loss": 0.
|
| 13 |
-
"training_worker_0_memory_MB":
|
| 14 |
-
"training_gpu_0_memory_MB":
|
| 15 |
-
"validation_start_acc": 0.
|
| 16 |
"validation_end_acc": 0.6535477767265847,
|
| 17 |
-
"validation_span_acc": 0.
|
| 18 |
-
"validation_em": 0.
|
| 19 |
-
"validation_f1": 0.
|
| 20 |
-
"validation_loss": 3.
|
| 21 |
-
"best_validation_start_acc": 0.
|
| 22 |
-
"best_validation_end_acc": 0.
|
| 23 |
-
"best_validation_span_acc": 0.
|
| 24 |
-
"best_validation_em": 0.
|
| 25 |
-
"best_validation_f1": 0.
|
| 26 |
-
"best_validation_loss": 2.
|
| 27 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_epoch": 10,
|
| 3 |
+
"peak_worker_0_memory_MB": 10196.45703125,
|
| 4 |
"peak_gpu_0_memory_MB": 2483.1318359375,
|
| 5 |
+
"training_duration": "1:17:24.504377",
|
| 6 |
+
"epoch": 19,
|
| 7 |
+
"training_start_acc": 0.8321213712485303,
|
| 8 |
+
"training_end_acc": 0.8789598054772315,
|
| 9 |
+
"training_span_acc": 0.7638899987442779,
|
| 10 |
+
"training_em": 0.8128631605383623,
|
| 11 |
+
"training_f1": 0.9108974145873597,
|
| 12 |
+
"training_loss": 0.8540654097245708,
|
| 13 |
+
"training_worker_0_memory_MB": 10196.45703125,
|
| 14 |
+
"training_gpu_0_memory_MB": 2470.0771484375,
|
| 15 |
+
"validation_start_acc": 0.6192999053926206,
|
| 16 |
"validation_end_acc": 0.6535477767265847,
|
| 17 |
+
"validation_span_acc": 0.528003784295175,
|
| 18 |
+
"validation_em": 0.659508041627247,
|
| 19 |
+
"validation_f1": 0.7623738533824579,
|
| 20 |
+
"validation_loss": 3.530728134569132,
|
| 21 |
+
"best_validation_start_acc": 0.6322611163670766,
|
| 22 |
+
"best_validation_end_acc": 0.6654683065279092,
|
| 23 |
+
"best_validation_span_acc": 0.5429517502365184,
|
| 24 |
+
"best_validation_em": 0.6717123935666982,
|
| 25 |
+
"best_validation_f1": 0.7701806420178526,
|
| 26 |
+
"best_validation_loss": 2.784988238676539
|
| 27 |
}
|
vocabulary/non_padded_namespaces.txt
CHANGED
|
@@ -1,2 +1,2 @@
|
|
| 1 |
-
*labels
|
| 2 |
*tags
|
|
|
|
|
|
|
|
|
| 1 |
*tags
|
| 2 |
+
*labels
|
weights.th
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef195409cc6eabb0262f893a8a6b6766ed9e6f456ed8401d37893e376df29f61
|
| 3 |
+
size 49008050
|