Training in progress, step 1000
Browse files- config.json +40 -0
- runs/May27_08-14-17_ae63705f58eb/events.out.tfevents.1716797668.ae63705f58eb.66261.0 +3 -0
- runs/May27_08-15-50_ae63705f58eb/events.out.tfevents.1716797761.ae63705f58eb.66571.0 +3 -0
- runs/May27_08-17-51_ae63705f58eb/events.out.tfevents.1716797897.ae63705f58eb.66932.0 +3 -0
- runs/May27_08-24-55_ae63705f58eb/events.out.tfevents.1716798313.ae63705f58eb.67678.0 +3 -0
- training_args.bin +3 -0
config.json
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/paligemma-3b-pt-224",
|
3 |
+
"architectures": [
|
4 |
+
"PaliGemmaForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"bos_token_id": 2,
|
7 |
+
"eos_token_id": 1,
|
8 |
+
"hidden_size": 2048,
|
9 |
+
"ignore_index": -100,
|
10 |
+
"image_token_index": 257152,
|
11 |
+
"model_type": "paligemma",
|
12 |
+
"pad_token_id": 0,
|
13 |
+
"projection_dim": 2048,
|
14 |
+
"text_config": {
|
15 |
+
"hidden_size": 2048,
|
16 |
+
"intermediate_size": 16384,
|
17 |
+
"model_type": "gemma",
|
18 |
+
"num_attention_heads": 8,
|
19 |
+
"num_hidden_layers": 18,
|
20 |
+
"num_image_tokens": 256,
|
21 |
+
"num_key_value_heads": 1,
|
22 |
+
"torch_dtype": "float32",
|
23 |
+
"vocab_size": 257216
|
24 |
+
},
|
25 |
+
"torch_dtype": "bfloat16",
|
26 |
+
"transformers_version": "4.41.0",
|
27 |
+
"vision_config": {
|
28 |
+
"hidden_size": 1152,
|
29 |
+
"intermediate_size": 4304,
|
30 |
+
"model_type": "siglip_vision_model",
|
31 |
+
"num_attention_heads": 16,
|
32 |
+
"num_hidden_layers": 27,
|
33 |
+
"num_image_tokens": 256,
|
34 |
+
"patch_size": 14,
|
35 |
+
"projection_dim": 2048,
|
36 |
+
"projector_hidden_act": "gelu_fast",
|
37 |
+
"vision_use_head": false
|
38 |
+
},
|
39 |
+
"vocab_size": 257216
|
40 |
+
}
|
runs/May27_08-14-17_ae63705f58eb/events.out.tfevents.1716797668.ae63705f58eb.66261.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0fca425621067e3dbe7d71b1e0ea50440342728c86d0c6cb8c9fd122954cee84
|
3 |
+
size 5153
|
runs/May27_08-15-50_ae63705f58eb/events.out.tfevents.1716797761.ae63705f58eb.66571.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:46d87269bd6374f04ac38011f194d1ed6da1f9547bba04aaa542324e0a284d57
|
3 |
+
size 5153
|
runs/May27_08-17-51_ae63705f58eb/events.out.tfevents.1716797897.ae63705f58eb.66932.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e014e6ab1bf8022c0f30ef2ca74dfb3c24210b118ceb40d421742e942ed24d12
|
3 |
+
size 5153
|
runs/May27_08-24-55_ae63705f58eb/events.out.tfevents.1716798313.ae63705f58eb.67678.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:43041a592a3d58a31cb79ca4360e9f365bb0a3abd26838d5aa0de82c542bd208
|
3 |
+
size 7801
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0bed9420d7f5f3d371e94f0291b4afd2ae566e37abc92a6eb5f4cde492bf0d0e
|
3 |
+
size 5112
|