Training in progress, epoch 1
Browse files
config.json
CHANGED
|
@@ -8,7 +8,7 @@
|
|
| 8 |
"initializer_factor": 1.0,
|
| 9 |
"initializer_range": 0.02,
|
| 10 |
"is_encoder_decoder": true,
|
| 11 |
-
"is_vqa":
|
| 12 |
"model_type": "pix2struct",
|
| 13 |
"pad_token_id": 0,
|
| 14 |
"text_config": {
|
|
@@ -19,7 +19,7 @@
|
|
| 19 |
"d_kv": 64,
|
| 20 |
"decoder_start_token_id": 0,
|
| 21 |
"dense_act_fn": "gelu_new",
|
| 22 |
-
"dropout_rate": 0.
|
| 23 |
"dtype": "float32",
|
| 24 |
"encoder_hidden_size": 768,
|
| 25 |
"eos_token_id": 1,
|
|
@@ -54,17 +54,18 @@
|
|
| 54 |
"use_cache": false,
|
| 55 |
"vision_config": {
|
| 56 |
"add_cross_attention": false,
|
| 57 |
-
"attention_dropout": 0.
|
| 58 |
"bos_token_id": null,
|
| 59 |
"cross_attention_hidden_size": null,
|
| 60 |
"d_ff": 2048,
|
| 61 |
"d_kv": 64,
|
| 62 |
"decoder_start_token_id": null,
|
| 63 |
"dense_act_fn": "gelu_new",
|
| 64 |
-
"dropout_rate": 0.
|
| 65 |
"dtype": "float32",
|
| 66 |
"eos_token_id": null,
|
| 67 |
"finetuning_task": null,
|
|
|
|
| 68 |
"hidden_size": 768,
|
| 69 |
"initializer_factor": 1.0,
|
| 70 |
"initializer_range": 0.02,
|
|
|
|
| 8 |
"initializer_factor": 1.0,
|
| 9 |
"initializer_range": 0.02,
|
| 10 |
"is_encoder_decoder": true,
|
| 11 |
+
"is_vqa": false,
|
| 12 |
"model_type": "pix2struct",
|
| 13 |
"pad_token_id": 0,
|
| 14 |
"text_config": {
|
|
|
|
| 19 |
"d_kv": 64,
|
| 20 |
"decoder_start_token_id": 0,
|
| 21 |
"dense_act_fn": "gelu_new",
|
| 22 |
+
"dropout_rate": 0.2,
|
| 23 |
"dtype": "float32",
|
| 24 |
"encoder_hidden_size": 768,
|
| 25 |
"eos_token_id": 1,
|
|
|
|
| 54 |
"use_cache": false,
|
| 55 |
"vision_config": {
|
| 56 |
"add_cross_attention": false,
|
| 57 |
+
"attention_dropout": 0.2,
|
| 58 |
"bos_token_id": null,
|
| 59 |
"cross_attention_hidden_size": null,
|
| 60 |
"d_ff": 2048,
|
| 61 |
"d_kv": 64,
|
| 62 |
"decoder_start_token_id": null,
|
| 63 |
"dense_act_fn": "gelu_new",
|
| 64 |
+
"dropout_rate": 0.2,
|
| 65 |
"dtype": "float32",
|
| 66 |
"eos_token_id": null,
|
| 67 |
"finetuning_task": null,
|
| 68 |
+
"hidden_dropout_prob": 0.2,
|
| 69 |
"hidden_size": 768,
|
| 70 |
"initializer_factor": 1.0,
|
| 71 |
"initializer_range": 0.02,
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1130333048
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c5d137cd245ecfc4d52901210ab3114e43c99ee5656d28313608b7537551e72e
|
| 3 |
size 1130333048
|
runs/Mar25_20-21-43_81ab0780e482/events.out.tfevents.1774470103.81ab0780e482.673.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:965b08f492d899ef18e37ff225f2547bd5694f83c1863bf1cf5ff9d04fb398c7
|
| 3 |
+
size 6465
|
runs/Mar25_20-22-14_81ab0780e482/events.out.tfevents.1774470134.81ab0780e482.673.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5935768c60c4dafcf2cf3d83f314dd2be55322dc676b7e82fca5901368ffa81
|
| 3 |
+
size 19493
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5393
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:59d76d7e312c2383badd1a78bc4c067c5a0ffff58e032384d8a5361498122fb5
|
| 3 |
size 5393
|