- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/config.json +98 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/generation_config.json +7 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model-00001-of-00003.safetensors +3 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model-00002-of-00003.safetensors +3 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model-00003-of-00003.safetensors +3 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model.safetensors.index.json +0 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/trainer_state.json +485 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/training_args.bin +3 -0
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/config.json
ADDED
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"LlavaInterleaveForConditionalGeneration"
|
4 |
+
],
|
5 |
+
"btadapter": true,
|
6 |
+
"btadapter_depth": 4,
|
7 |
+
"clip_post_pretrain": null,
|
8 |
+
"clip_weight": "/group/40043/ruyangliu/weight/CLIP-L334",
|
9 |
+
"frame_shape": [
|
10 |
+
24,
|
11 |
+
24
|
12 |
+
],
|
13 |
+
"hidden_size": 4096,
|
14 |
+
"ignore_index": -100,
|
15 |
+
"image_grid_pinpoints": [
|
16 |
+
[
|
17 |
+
336,
|
18 |
+
672
|
19 |
+
],
|
20 |
+
[
|
21 |
+
672,
|
22 |
+
336
|
23 |
+
],
|
24 |
+
[
|
25 |
+
672,
|
26 |
+
672
|
27 |
+
],
|
28 |
+
[
|
29 |
+
1008,
|
30 |
+
336
|
31 |
+
],
|
32 |
+
[
|
33 |
+
336,
|
34 |
+
1008
|
35 |
+
]
|
36 |
+
],
|
37 |
+
"image_pooling_kernel": [
|
38 |
+
1,
|
39 |
+
3,
|
40 |
+
3
|
41 |
+
],
|
42 |
+
"image_pooling_stride": [
|
43 |
+
1,
|
44 |
+
3,
|
45 |
+
3
|
46 |
+
],
|
47 |
+
"image_token_index": 32000,
|
48 |
+
"long_clip": true,
|
49 |
+
"max_T": 64,
|
50 |
+
"model_type": "llava_next",
|
51 |
+
"pad_token_id": 0,
|
52 |
+
"pllava_pooling_shape": null,
|
53 |
+
"pooling": "clipST_3d",
|
54 |
+
"pooling_kernel": [
|
55 |
+
2,
|
56 |
+
3,
|
57 |
+
3
|
58 |
+
],
|
59 |
+
"pooling_stride": [
|
60 |
+
2,
|
61 |
+
3,
|
62 |
+
3
|
63 |
+
],
|
64 |
+
"pooling_temp": 0.01,
|
65 |
+
"projector_hidden_act": "gelu",
|
66 |
+
"text_config": {
|
67 |
+
"_name_or_path": "lmsys/vicuna-7b-v1.5",
|
68 |
+
"architectures": [
|
69 |
+
"LlamaForCausalLM"
|
70 |
+
],
|
71 |
+
"max_position_embeddings": 4096,
|
72 |
+
"model_type": "llama",
|
73 |
+
"pad_token_id": 0,
|
74 |
+
"rms_norm_eps": 1e-05,
|
75 |
+
"torch_dtype": "float16",
|
76 |
+
"vocab_size": 32064
|
77 |
+
},
|
78 |
+
"torch_dtype": "bfloat16",
|
79 |
+
"transformers_version": "4.40.0",
|
80 |
+
"use_image_newline_parameter": true,
|
81 |
+
"video_input": "mean",
|
82 |
+
"vision_config": {
|
83 |
+
"depth": 4,
|
84 |
+
"hidden_size": 1024,
|
85 |
+
"image_size": 336,
|
86 |
+
"intermediate_size": 4096,
|
87 |
+
"max_T": 64,
|
88 |
+
"model_type": "clip_vision_model",
|
89 |
+
"num_attention_heads": 16,
|
90 |
+
"num_hidden_layers": 24,
|
91 |
+
"patch_size": 14,
|
92 |
+
"projection_dim": 768,
|
93 |
+
"vocab_size": 32000
|
94 |
+
},
|
95 |
+
"vision_feature_layer": -2,
|
96 |
+
"vision_feature_select_strategy": "default",
|
97 |
+
"vocab_size": 32064
|
98 |
+
}
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.40.0"
|
7 |
+
}
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model-00001-of-00003.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2468064fc5ed3bc395743532b2d6565ee20a678efb5a292527aa25d80a547574
|
3 |
+
size 4980641554
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model-00002-of-00003.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6093c7766f8f45b2f51be4ebb77ed2df5c49b410970e0620a6d5d4ff7e8b0c13
|
3 |
+
size 4957878536
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model-00003-of-00003.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e74b0d8a25e3c65d3965fcfdaf83dd54ef2f8087c14d12f202a80af25e0ad167
|
3 |
+
size 4580905400
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model.safetensors.index.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/trainer_state.json
ADDED
@@ -0,0 +1,485 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.9998461775111521,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 3250,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.015382248884786957,
|
13 |
+
"grad_norm": 1.2864877128592522,
|
14 |
+
"learning_rate": 1.0204081632653063e-05,
|
15 |
+
"loss": 1.211,
|
16 |
+
"step": 50
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.030764497769573913,
|
20 |
+
"grad_norm": 1.227813083166002,
|
21 |
+
"learning_rate": 1.999998013186146e-05,
|
22 |
+
"loss": 1.0177,
|
23 |
+
"step": 100
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.046146746654360866,
|
27 |
+
"grad_norm": 1.16006238153043,
|
28 |
+
"learning_rate": 1.9986572140101312e-05,
|
29 |
+
"loss": 0.9909,
|
30 |
+
"step": 150
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.06152899553914783,
|
34 |
+
"grad_norm": 1.2120743368126807,
|
35 |
+
"learning_rate": 1.9948367447837503e-05,
|
36 |
+
"loss": 0.9853,
|
37 |
+
"step": 200
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.07691124442393478,
|
41 |
+
"grad_norm": 1.0434928463049529,
|
42 |
+
"learning_rate": 1.9885460917481137e-05,
|
43 |
+
"loss": 0.9766,
|
44 |
+
"step": 250
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.09229349330872173,
|
48 |
+
"grad_norm": 1.0799628915416184,
|
49 |
+
"learning_rate": 1.9798008746210894e-05,
|
50 |
+
"loss": 0.9819,
|
51 |
+
"step": 300
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.10767574219350869,
|
55 |
+
"grad_norm": 1.0390775292248948,
|
56 |
+
"learning_rate": 1.968622807813479e-05,
|
57 |
+
"loss": 0.9678,
|
58 |
+
"step": 350
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.12305799107829565,
|
62 |
+
"grad_norm": 1.3974223368156584,
|
63 |
+
"learning_rate": 1.955039646512041e-05,
|
64 |
+
"loss": 0.9782,
|
65 |
+
"step": 400
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.1384402399630826,
|
69 |
+
"grad_norm": 1.1107339192286112,
|
70 |
+
"learning_rate": 1.9390851177632496e-05,
|
71 |
+
"loss": 0.9602,
|
72 |
+
"step": 450
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.15382248884786956,
|
76 |
+
"grad_norm": 1.0579993349228267,
|
77 |
+
"learning_rate": 1.9207988367288956e-05,
|
78 |
+
"loss": 0.9571,
|
79 |
+
"step": 500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.1692047377326565,
|
83 |
+
"grad_norm": 1.1144151233888806,
|
84 |
+
"learning_rate": 1.9002262083214728e-05,
|
85 |
+
"loss": 0.9669,
|
86 |
+
"step": 550
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.18458698661744347,
|
90 |
+
"grad_norm": 1.187591950721483,
|
91 |
+
"learning_rate": 1.8774183144635894e-05,
|
92 |
+
"loss": 0.958,
|
93 |
+
"step": 600
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.19996923550223042,
|
97 |
+
"grad_norm": 1.3767826488829,
|
98 |
+
"learning_rate": 1.8524317872513434e-05,
|
99 |
+
"loss": 0.9481,
|
100 |
+
"step": 650
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.21535148438701737,
|
104 |
+
"grad_norm": 1.0853939205331895,
|
105 |
+
"learning_rate": 1.8253286683365855e-05,
|
106 |
+
"loss": 0.9434,
|
107 |
+
"step": 700
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.23073373327180433,
|
111 |
+
"grad_norm": 1.0358359930096885,
|
112 |
+
"learning_rate": 1.7961762548772413e-05,
|
113 |
+
"loss": 0.9524,
|
114 |
+
"step": 750
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.2461159821565913,
|
118 |
+
"grad_norm": 1.0885430781645802,
|
119 |
+
"learning_rate": 1.7650469324381843e-05,
|
120 |
+
"loss": 0.9443,
|
121 |
+
"step": 800
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.26149823104137826,
|
125 |
+
"grad_norm": 1.139432989706956,
|
126 |
+
"learning_rate": 1.732017995257575e-05,
|
127 |
+
"loss": 0.9398,
|
128 |
+
"step": 850
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.2768804799261652,
|
132 |
+
"grad_norm": 1.1616765089635157,
|
133 |
+
"learning_rate": 1.697171454324943e-05,
|
134 |
+
"loss": 0.9318,
|
135 |
+
"step": 900
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.29226272881095217,
|
139 |
+
"grad_norm": 1.102418985915195,
|
140 |
+
"learning_rate": 1.660593833747557e-05,
|
141 |
+
"loss": 0.9348,
|
142 |
+
"step": 950
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.3076449776957391,
|
146 |
+
"grad_norm": 1.0377834162761999,
|
147 |
+
"learning_rate": 1.6223759559107028e-05,
|
148 |
+
"loss": 0.9206,
|
149 |
+
"step": 1000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.3230272265805261,
|
153 |
+
"grad_norm": 1.103549274843009,
|
154 |
+
"learning_rate": 1.5826127159653203e-05,
|
155 |
+
"loss": 0.9353,
|
156 |
+
"step": 1050
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.338409475465313,
|
160 |
+
"grad_norm": 1.0090406165137933,
|
161 |
+
"learning_rate": 1.5414028462029454e-05,
|
162 |
+
"loss": 0.9363,
|
163 |
+
"step": 1100
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.3537917243501,
|
167 |
+
"grad_norm": 1.1329344585144494,
|
168 |
+
"learning_rate": 1.4988486709030166e-05,
|
169 |
+
"loss": 0.9347,
|
170 |
+
"step": 1150
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.36917397323488693,
|
174 |
+
"grad_norm": 1.0307615452725163,
|
175 |
+
"learning_rate": 1.4550558522612584e-05,
|
176 |
+
"loss": 0.9231,
|
177 |
+
"step": 1200
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.3845562221196739,
|
181 |
+
"grad_norm": 1.1133951229282608,
|
182 |
+
"learning_rate": 1.410133128030009e-05,
|
183 |
+
"loss": 0.9281,
|
184 |
+
"step": 1250
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.39993847100446084,
|
188 |
+
"grad_norm": 1.0490189419573543,
|
189 |
+
"learning_rate": 1.3641920415219224e-05,
|
190 |
+
"loss": 0.9229,
|
191 |
+
"step": 1300
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.4153207198892478,
|
195 |
+
"grad_norm": 1.0225887710538208,
|
196 |
+
"learning_rate": 1.3173466646474575e-05,
|
197 |
+
"loss": 0.9115,
|
198 |
+
"step": 1350
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.43070296877403474,
|
202 |
+
"grad_norm": 1.0818630597333305,
|
203 |
+
"learning_rate": 1.2697133146738452e-05,
|
204 |
+
"loss": 0.9124,
|
205 |
+
"step": 1400
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.4460852176588217,
|
209 |
+
"grad_norm": 1.1222485167556653,
|
210 |
+
"learning_rate": 1.2214102654088248e-05,
|
211 |
+
"loss": 0.9166,
|
212 |
+
"step": 1450
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.46146746654360865,
|
216 |
+
"grad_norm": 0.9980409625749838,
|
217 |
+
"learning_rate": 1.1725574535262863e-05,
|
218 |
+
"loss": 0.9047,
|
219 |
+
"step": 1500
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.4768497154283956,
|
223 |
+
"grad_norm": 1.0397039742249545,
|
224 |
+
"learning_rate": 1.1232761807630032e-05,
|
225 |
+
"loss": 0.9079,
|
226 |
+
"step": 1550
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.4922319643131826,
|
230 |
+
"grad_norm": 1.0715968646982938,
|
231 |
+
"learning_rate": 1.0736888127259132e-05,
|
232 |
+
"loss": 0.9061,
|
233 |
+
"step": 1600
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.5076142131979695,
|
237 |
+
"grad_norm": 1.0724763367894545,
|
238 |
+
"learning_rate": 1.023918475057803e-05,
|
239 |
+
"loss": 0.8947,
|
240 |
+
"step": 1650
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.5229964620827565,
|
244 |
+
"grad_norm": 1.03882946183293,
|
245 |
+
"learning_rate": 9.740887477158194e-06,
|
246 |
+
"loss": 0.9053,
|
247 |
+
"step": 1700
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.5383787109675434,
|
251 |
+
"grad_norm": 1.0204487030578802,
|
252 |
+
"learning_rate": 9.24323358121919e-06,
|
253 |
+
"loss": 0.9016,
|
254 |
+
"step": 1750
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.5537609598523304,
|
258 |
+
"grad_norm": 1.0085542891249055,
|
259 |
+
"learning_rate": 8.747458739471603e-06,
|
260 |
+
"loss": 0.8988,
|
261 |
+
"step": 1800
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.5691432087371173,
|
265 |
+
"grad_norm": 1.1293770623412562,
|
266 |
+
"learning_rate": 8.254793962926579e-06,
|
267 |
+
"loss": 0.8967,
|
268 |
+
"step": 1850
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 0.5845254576219043,
|
272 |
+
"grad_norm": 1.1860769177336596,
|
273 |
+
"learning_rate": 7.76646254029032e-06,
|
274 |
+
"loss": 0.8967,
|
275 |
+
"step": 1900
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 0.5999077065066912,
|
279 |
+
"grad_norm": 1.062132771926321,
|
280 |
+
"learning_rate": 7.283677000533106e-06,
|
281 |
+
"loss": 0.8936,
|
282 |
+
"step": 1950
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 0.6152899553914782,
|
286 |
+
"grad_norm": 1.1679358844534626,
|
287 |
+
"learning_rate": 6.807636102174764e-06,
|
288 |
+
"loss": 0.8928,
|
289 |
+
"step": 2000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 0.6306722042762651,
|
293 |
+
"grad_norm": 1.0610336394985995,
|
294 |
+
"learning_rate": 6.339521856762254e-06,
|
295 |
+
"loss": 0.8845,
|
296 |
+
"step": 2050
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.6460544531610521,
|
300 |
+
"grad_norm": 1.084745782348564,
|
301 |
+
"learning_rate": 5.880496593930021e-06,
|
302 |
+
"loss": 0.8869,
|
303 |
+
"step": 2100
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 0.6614367020458392,
|
307 |
+
"grad_norm": 0.9948936563241075,
|
308 |
+
"learning_rate": 5.431700075330654e-06,
|
309 |
+
"loss": 0.8817,
|
310 |
+
"step": 2150
|
311 |
+
},
|
312 |
+
{
|
313 |
+
"epoch": 0.676818950930626,
|
314 |
+
"grad_norm": 1.0955833872820746,
|
315 |
+
"learning_rate": 4.9942466646019e-06,
|
316 |
+
"loss": 0.8857,
|
317 |
+
"step": 2200
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 0.6922011998154131,
|
321 |
+
"grad_norm": 1.0328397346172227,
|
322 |
+
"learning_rate": 4.569222560397126e-06,
|
323 |
+
"loss": 0.8842,
|
324 |
+
"step": 2250
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 0.7075834487002,
|
328 |
+
"grad_norm": 1.0439140807601641,
|
329 |
+
"learning_rate": 4.157683099349444e-06,
|
330 |
+
"loss": 0.8797,
|
331 |
+
"step": 2300
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 0.722965697584987,
|
335 |
+
"grad_norm": 1.0659188075287298,
|
336 |
+
"learning_rate": 3.760650135666476e-06,
|
337 |
+
"loss": 0.8765,
|
338 |
+
"step": 2350
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.7383479464697739,
|
342 |
+
"grad_norm": 1.0691950735737183,
|
343 |
+
"learning_rate": 3.3791095038620515e-06,
|
344 |
+
"loss": 0.8715,
|
345 |
+
"step": 2400
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 0.7537301953545609,
|
349 |
+
"grad_norm": 1.0882575411106976,
|
350 |
+
"learning_rate": 3.0140085709249666e-06,
|
351 |
+
"loss": 0.8754,
|
352 |
+
"step": 2450
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 0.7691124442393478,
|
356 |
+
"grad_norm": 1.1112353348166593,
|
357 |
+
"learning_rate": 2.6662538840027984e-06,
|
358 |
+
"loss": 0.873,
|
359 |
+
"step": 2500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 0.7844946931241348,
|
363 |
+
"grad_norm": 1.1312531360561675,
|
364 |
+
"learning_rate": 2.3367089194415226e-06,
|
365 |
+
"loss": 0.8776,
|
366 |
+
"step": 2550
|
367 |
+
},
|
368 |
+
{
|
369 |
+
"epoch": 0.7998769420089217,
|
370 |
+
"grad_norm": 1.118822941453432,
|
371 |
+
"learning_rate": 2.026191938770128e-06,
|
372 |
+
"loss": 0.8769,
|
373 |
+
"step": 2600
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 0.8152591908937087,
|
377 |
+
"grad_norm": 1.0617562102838363,
|
378 |
+
"learning_rate": 1.7354739569538081e-06,
|
379 |
+
"loss": 0.8611,
|
380 |
+
"step": 2650
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.8306414397784956,
|
384 |
+
"grad_norm": 1.1190891753348724,
|
385 |
+
"learning_rate": 1.4652768279605734e-06,
|
386 |
+
"loss": 0.8778,
|
387 |
+
"step": 2700
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.8460236886632826,
|
391 |
+
"grad_norm": 1.1025665252752193,
|
392 |
+
"learning_rate": 1.2162714523948104e-06,
|
393 |
+
"loss": 0.8582,
|
394 |
+
"step": 2750
|
395 |
+
},
|
396 |
+
{
|
397 |
+
"epoch": 0.8614059375480695,
|
398 |
+
"grad_norm": 1.0621744666965152,
|
399 |
+
"learning_rate": 9.89076111648284e-07,
|
400 |
+
"loss": 0.8643,
|
401 |
+
"step": 2800
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 0.8767881864328565,
|
405 |
+
"grad_norm": 0.9763313630276178,
|
406 |
+
"learning_rate": 7.842549327048366e-07,
|
407 |
+
"loss": 0.8667,
|
408 |
+
"step": 2850
|
409 |
+
},
|
410 |
+
{
|
411 |
+
"epoch": 0.8921704353176434,
|
412 |
+
"grad_norm": 1.0393762329975094,
|
413 |
+
"learning_rate": 6.023164874107202e-07,
|
414 |
+
"loss": 0.8677,
|
415 |
+
"step": 2900
|
416 |
+
},
|
417 |
+
{
|
418 |
+
"epoch": 0.9075526842024304,
|
419 |
+
"grad_norm": 1.1026189621389317,
|
420 |
+
"learning_rate": 4.437125296885847e-07,
|
421 |
+
"loss": 0.8649,
|
422 |
+
"step": 2950
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 0.9229349330872173,
|
426 |
+
"grad_norm": 1.073850367410434,
|
427 |
+
"learning_rate": 3.088368738305747e-07,
|
428 |
+
"loss": 0.8771,
|
429 |
+
"step": 3000
|
430 |
+
},
|
431 |
+
{
|
432 |
+
"epoch": 0.9383171819720043,
|
433 |
+
"grad_norm": 1.0327578416174346,
|
434 |
+
"learning_rate": 1.9802441665580208e-07,
|
435 |
+
"loss": 0.8678,
|
436 |
+
"step": 3050
|
437 |
+
},
|
438 |
+
{
|
439 |
+
"epoch": 0.9536994308567912,
|
440 |
+
"grad_norm": 1.035131432577654,
|
441 |
+
"learning_rate": 1.1155030596013638e-07,
|
442 |
+
"loss": 0.8662,
|
443 |
+
"step": 3100
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 0.9690816797415782,
|
447 |
+
"grad_norm": 1.0361099744000193,
|
448 |
+
"learning_rate": 4.962925732307278e-08,
|
449 |
+
"loss": 0.8588,
|
450 |
+
"step": 3150
|
451 |
+
},
|
452 |
+
{
|
453 |
+
"epoch": 0.9844639286263652,
|
454 |
+
"grad_norm": 1.087919608094763,
|
455 |
+
"learning_rate": 1.2415020968049585e-08,
|
456 |
+
"loss": 0.8733,
|
457 |
+
"step": 3200
|
458 |
+
},
|
459 |
+
{
|
460 |
+
"epoch": 0.9998461775111521,
|
461 |
+
"grad_norm": 1.0473814351391442,
|
462 |
+
"learning_rate": 0.0,
|
463 |
+
"loss": 0.8751,
|
464 |
+
"step": 3250
|
465 |
+
},
|
466 |
+
{
|
467 |
+
"epoch": 0.9998461775111521,
|
468 |
+
"step": 3250,
|
469 |
+
"total_flos": 1.7053270040117248e+16,
|
470 |
+
"train_loss": 0.9156783118614784,
|
471 |
+
"train_runtime": 46310.2792,
|
472 |
+
"train_samples_per_second": 0.562,
|
473 |
+
"train_steps_per_second": 0.07
|
474 |
+
}
|
475 |
+
],
|
476 |
+
"logging_steps": 50,
|
477 |
+
"max_steps": 3250,
|
478 |
+
"num_input_tokens_seen": 0,
|
479 |
+
"num_train_epochs": 1,
|
480 |
+
"save_steps": 500,
|
481 |
+
"total_flos": 1.7053270040117248e+16,
|
482 |
+
"train_batch_size": 4,
|
483 |
+
"trial_name": null,
|
484 |
+
"trial_params": null
|
485 |
+
}
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3db21088a03dab3e9042367b9f7e3bb67dfc608a5563e28fb3fb76fa7ab52d6c
|
3 |
+
size 7032
|