Commit
·
c004e9c
1
Parent(s):
06b296d
Training in progress, step 500
Browse files- config.json +46 -0
- gaudi_config.json +33 -0
- pytorch_model.bin +3 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "BridgeTower/bridgetower-large-itm-mlm-itc",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BridgeTowerForContrastiveLearning"
|
| 5 |
+
],
|
| 6 |
+
"contrastive_hidden_size": 512,
|
| 7 |
+
"drop_rate": 0.1,
|
| 8 |
+
"head_hidden_scale": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_size": 1024,
|
| 11 |
+
"init_layernorm_from_vision_encoder": false,
|
| 12 |
+
"initializer_factor": 1,
|
| 13 |
+
"layer_norm_eps": 1e-05,
|
| 14 |
+
"link_tower_type": "add",
|
| 15 |
+
"logit_scale_init_value": 2.6592,
|
| 16 |
+
"model_type": "bridgetower",
|
| 17 |
+
"num_attention_heads": 16,
|
| 18 |
+
"num_hidden_layers": 6,
|
| 19 |
+
"share_cross_modal_transformer_layers": true,
|
| 20 |
+
"share_link_tower_layers": false,
|
| 21 |
+
"text_config": {
|
| 22 |
+
"architectures": [
|
| 23 |
+
"BridgeTowerTextModel"
|
| 24 |
+
],
|
| 25 |
+
"classifier_dropout": null,
|
| 26 |
+
"hidden_size": 1024,
|
| 27 |
+
"initializer_range": 0.02,
|
| 28 |
+
"intermediate_size": 4096,
|
| 29 |
+
"model_type": "bridgetower_text_model",
|
| 30 |
+
"num_attention_heads": 16,
|
| 31 |
+
"num_hidden_layers": 24
|
| 32 |
+
},
|
| 33 |
+
"tie_word_embeddings": false,
|
| 34 |
+
"torch_dtype": "float32",
|
| 35 |
+
"transformers_version": "4.34.1",
|
| 36 |
+
"vision_config": {
|
| 37 |
+
"architectures": [
|
| 38 |
+
"BridgeTowerVisionModel"
|
| 39 |
+
],
|
| 40 |
+
"hidden_size": 1024,
|
| 41 |
+
"image_size": 294,
|
| 42 |
+
"model_type": "bridgetower_vision_model",
|
| 43 |
+
"num_hidden_layers": 24,
|
| 44 |
+
"patch_size": 14
|
| 45 |
+
}
|
| 46 |
+
}
|
gaudi_config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"autocast_bf16_ops": null,
|
| 3 |
+
"autocast_fp32_ops": null,
|
| 4 |
+
"hmp_bf16_ops": [
|
| 5 |
+
"add",
|
| 6 |
+
"addmm",
|
| 7 |
+
"bmm",
|
| 8 |
+
"div",
|
| 9 |
+
"dropout",
|
| 10 |
+
"gelu",
|
| 11 |
+
"iadd",
|
| 12 |
+
"linear",
|
| 13 |
+
"layer_norm",
|
| 14 |
+
"matmul",
|
| 15 |
+
"mm",
|
| 16 |
+
"rsub",
|
| 17 |
+
"softmax",
|
| 18 |
+
"truediv"
|
| 19 |
+
],
|
| 20 |
+
"hmp_fp32_ops": [
|
| 21 |
+
"embedding",
|
| 22 |
+
"nll_loss",
|
| 23 |
+
"log_softmax"
|
| 24 |
+
],
|
| 25 |
+
"hmp_is_verbose": false,
|
| 26 |
+
"hmp_opt_level": "O1",
|
| 27 |
+
"optimum_version": "1.8.0",
|
| 28 |
+
"transformers_version": "4.34.1",
|
| 29 |
+
"use_fused_adam": true,
|
| 30 |
+
"use_fused_clip_norm": true,
|
| 31 |
+
"use_habana_mixed_precision": false,
|
| 32 |
+
"use_torch_autocast": true
|
| 33 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7c4b558788f08343aff6ce1a1a3e9c147810805f9acfd79d41f03a8d9ebbd0d
|
| 3 |
+
size 3462613729
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85d2157d40df87abb135bd72f54f1f621dfb2268663545d517c265830ac98077
|
| 3 |
+
size 4539
|