graphcore-rahult
commited on
Commit
•
78fc866
1
Parent(s):
a283d19
Training in progress, step 500
Browse files- config.json +1 -1
- ipu_config.json +1 -1
- pytorch_model.bin +1 -1
- training_args.bin +1 -1
config.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
{
|
2 |
"_name_or_path": "roberta-base",
|
3 |
"architectures": [
|
4 |
-
"
|
5 |
],
|
6 |
"attention_probs_dropout_prob": 0.1,
|
7 |
"bos_token_id": 0,
|
|
|
1 |
{
|
2 |
"_name_or_path": "roberta-base",
|
3 |
"architectures": [
|
4 |
+
"PoptorchPipelinedRobertaForMaskedLM"
|
5 |
],
|
6 |
"attention_probs_dropout_prob": 0.1,
|
7 |
"bos_token_id": 0,
|
ipu_config.json
CHANGED
@@ -3,7 +3,7 @@
|
|
3 |
"device_iterations": 1,
|
4 |
"embedding_serialization_factor": 0,
|
5 |
"enable_half_partials": true,
|
6 |
-
"executable_cache_dir": "/tmp/exe_cache",
|
7 |
"execute_encoder_on_cpu_for_generation": false,
|
8 |
"gradient_accumulation_steps": 16,
|
9 |
"inference_device_iterations": 5,
|
|
|
3 |
"device_iterations": 1,
|
4 |
"embedding_serialization_factor": 0,
|
5 |
"enable_half_partials": true,
|
6 |
+
"executable_cache_dir": "/tmp/exe_cache/language_modeling",
|
7 |
"execute_encoder_on_cpu_for_generation": false,
|
8 |
"gradient_accumulation_steps": 16,
|
9 |
"inference_device_iterations": 5,
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 249463979
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c46b9dce35d1478d85575c0ca4ba90750a525885439356a8e3d59391161d0b3f
|
3 |
size 249463979
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2735
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:38f26e9bcc3eb42d144bb8dbc62729fc961b032ef8520da97700809f52d77fa0
|
3 |
size 2735
|