first
Browse files- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/config.json +97 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/generation_config.json +7 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model-00001-of-00003.safetensors +3 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model-00002-of-00003.safetensors +3 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model-00003-of-00003.safetensors +3 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model.safetensors.index.json +0 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/trainer_state.json +422 -0
- llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/training_args.bin +3 -0
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/config.json
ADDED
@@ -0,0 +1,97 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
+
"vision_config": {
|
82 |
+
"depth": 4,
|
83 |
+
"hidden_size": 1024,
|
84 |
+
"image_size": 336,
|
85 |
+
"intermediate_size": 4096,
|
86 |
+
"max_T": 64,
|
87 |
+
"model_type": "clip_vision_model",
|
88 |
+
"num_attention_heads": 16,
|
89 |
+
"num_hidden_layers": 24,
|
90 |
+
"patch_size": 14,
|
91 |
+
"projection_dim": 768,
|
92 |
+
"vocab_size": 32000
|
93 |
+
},
|
94 |
+
"vision_feature_layer": -2,
|
95 |
+
"vision_feature_select_strategy": "default",
|
96 |
+
"vocab_size": 32064
|
97 |
+
}
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_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_multiimg_trainllm_e1_bz512_32f1688/model-00001-of-00003.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:badbca0f8b6f276fcf9f0b6f35ad29370715f5d9e5bde69804ab78457e611ced
|
3 |
+
size 4980641554
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model-00002-of-00003.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c486b79a3c76b9a43cfd068b592bfc88cfac2e5564d2552c55b24b518581eb0d
|
3 |
+
size 4957878536
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model-00003-of-00003.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e4df79bed0e8f9f55e5457c7bc44022a71d5a4d7f5e4402b4da25a356401f2cf
|
3 |
+
size 4580905400
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_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_multiimg_trainllm_e1_bz512_32f1688/trainer_state.json
ADDED
@@ -0,0 +1,422 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.9999115122555526,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 2825,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.017697548889478807,
|
13 |
+
"grad_norm": 1.3246452072458292,
|
14 |
+
"learning_rate": 1.1764705882352942e-05,
|
15 |
+
"loss": 1.3482,
|
16 |
+
"step": 50
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.035395097778957614,
|
20 |
+
"grad_norm": 1.184986868355016,
|
21 |
+
"learning_rate": 1.9998521094455198e-05,
|
22 |
+
"loss": 1.0655,
|
23 |
+
"step": 100
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.053092646668436425,
|
27 |
+
"grad_norm": 0.9956602761532123,
|
28 |
+
"learning_rate": 1.9972241607451552e-05,
|
29 |
+
"loss": 1.0325,
|
30 |
+
"step": 150
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.07079019555791523,
|
34 |
+
"grad_norm": 0.8985264278021503,
|
35 |
+
"learning_rate": 1.9913196946839304e-05,
|
36 |
+
"loss": 1.0115,
|
37 |
+
"step": 200
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.08848774444739403,
|
41 |
+
"grad_norm": 0.9761536009178263,
|
42 |
+
"learning_rate": 1.9821581111985072e-05,
|
43 |
+
"loss": 1.0087,
|
44 |
+
"step": 250
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.10618529333687285,
|
48 |
+
"grad_norm": 0.9582085355410745,
|
49 |
+
"learning_rate": 1.9697695119335547e-05,
|
50 |
+
"loss": 0.9959,
|
51 |
+
"step": 300
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.12388284222635165,
|
55 |
+
"grad_norm": 0.9502507945345694,
|
56 |
+
"learning_rate": 1.954194601338651e-05,
|
57 |
+
"loss": 0.9832,
|
58 |
+
"step": 350
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.14158039111583046,
|
62 |
+
"grad_norm": 0.9312008953589136,
|
63 |
+
"learning_rate": 1.9354845529282042e-05,
|
64 |
+
"loss": 0.9838,
|
65 |
+
"step": 400
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.15927794000530926,
|
69 |
+
"grad_norm": 0.9921814925845447,
|
70 |
+
"learning_rate": 1.9137008411438213e-05,
|
71 |
+
"loss": 0.9743,
|
72 |
+
"step": 450
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.17697548889478806,
|
76 |
+
"grad_norm": 0.9946357977754318,
|
77 |
+
"learning_rate": 1.8889150393715627e-05,
|
78 |
+
"loss": 0.9674,
|
79 |
+
"step": 500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.19467303778426687,
|
83 |
+
"grad_norm": 0.936001151695521,
|
84 |
+
"learning_rate": 1.8612085847777215e-05,
|
85 |
+
"loss": 0.9679,
|
86 |
+
"step": 550
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.2123705866737457,
|
90 |
+
"grad_norm": 0.9739097270294552,
|
91 |
+
"learning_rate": 1.8306725107357933e-05,
|
92 |
+
"loss": 0.9622,
|
93 |
+
"step": 600
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.2300681355632245,
|
97 |
+
"grad_norm": 1.0799028416482208,
|
98 |
+
"learning_rate": 1.7974071477237887e-05,
|
99 |
+
"loss": 0.976,
|
100 |
+
"step": 650
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.2477656844527033,
|
104 |
+
"grad_norm": 0.9330950566244366,
|
105 |
+
"learning_rate": 1.7615217936746246e-05,
|
106 |
+
"loss": 0.9567,
|
107 |
+
"step": 700
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.2654632333421821,
|
111 |
+
"grad_norm": 0.8832817518420761,
|
112 |
+
"learning_rate": 1.7231343548627085e-05,
|
113 |
+
"loss": 0.9523,
|
114 |
+
"step": 750
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.2831607822316609,
|
118 |
+
"grad_norm": 0.9002124999100791,
|
119 |
+
"learning_rate": 1.6823709585066308e-05,
|
120 |
+
"loss": 0.9466,
|
121 |
+
"step": 800
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.30085833112113974,
|
125 |
+
"grad_norm": 0.9157647153217237,
|
126 |
+
"learning_rate": 1.6393655383608132e-05,
|
127 |
+
"loss": 0.9429,
|
128 |
+
"step": 850
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.3185558800106185,
|
132 |
+
"grad_norm": 5.464496113406752,
|
133 |
+
"learning_rate": 1.594259394657707e-05,
|
134 |
+
"loss": 0.9695,
|
135 |
+
"step": 900
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.33625342890009735,
|
139 |
+
"grad_norm": 0.9883051330334294,
|
140 |
+
"learning_rate": 1.5472007298464117e-05,
|
141 |
+
"loss": 0.944,
|
142 |
+
"step": 950
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.3539509777895761,
|
146 |
+
"grad_norm": 0.8832131211036971,
|
147 |
+
"learning_rate": 1.4983441616531152e-05,
|
148 |
+
"loss": 0.937,
|
149 |
+
"step": 1000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.37164852667905496,
|
153 |
+
"grad_norm": 0.9595043070379917,
|
154 |
+
"learning_rate": 1.4478502150632503e-05,
|
155 |
+
"loss": 0.9313,
|
156 |
+
"step": 1050
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.38934607556853373,
|
160 |
+
"grad_norm": 1.035782879874017,
|
161 |
+
"learning_rate": 1.3958847948945428e-05,
|
162 |
+
"loss": 0.9264,
|
163 |
+
"step": 1100
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.40704362445801257,
|
167 |
+
"grad_norm": 0.961077579119761,
|
168 |
+
"learning_rate": 1.3426186406938769e-05,
|
169 |
+
"loss": 0.924,
|
170 |
+
"step": 1150
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.4247411733474914,
|
174 |
+
"grad_norm": 0.882487785914029,
|
175 |
+
"learning_rate": 1.2882267657489908e-05,
|
176 |
+
"loss": 0.9226,
|
177 |
+
"step": 1200
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.44243872223697017,
|
181 |
+
"grad_norm": 0.8676230908883836,
|
182 |
+
"learning_rate": 1.2328878820582122e-05,
|
183 |
+
"loss": 0.9134,
|
184 |
+
"step": 1250
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.460136271126449,
|
188 |
+
"grad_norm": 0.8663581800411811,
|
189 |
+
"learning_rate": 1.1767838131475654e-05,
|
190 |
+
"loss": 0.917,
|
191 |
+
"step": 1300
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.4778338200159278,
|
195 |
+
"grad_norm": 1.1913367079153512,
|
196 |
+
"learning_rate": 1.1200988966645286e-05,
|
197 |
+
"loss": 0.9181,
|
198 |
+
"step": 1350
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.4955313689054066,
|
202 |
+
"grad_norm": 0.8563122357771151,
|
203 |
+
"learning_rate": 1.0630193787112994e-05,
|
204 |
+
"loss": 0.903,
|
205 |
+
"step": 1400
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.5132289177948854,
|
209 |
+
"grad_norm": 0.8518715044091224,
|
210 |
+
"learning_rate": 1.005732801907567e-05,
|
211 |
+
"loss": 0.9085,
|
212 |
+
"step": 1450
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.5309264666843642,
|
216 |
+
"grad_norm": 0.9224776221871871,
|
217 |
+
"learning_rate": 9.484273891933982e-06,
|
218 |
+
"loss": 0.9086,
|
219 |
+
"step": 1500
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.548624015573843,
|
223 |
+
"grad_norm": 0.9846231662432429,
|
224 |
+
"learning_rate": 8.912914253968391e-06,
|
225 |
+
"loss": 0.9017,
|
226 |
+
"step": 1550
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.5663215644633218,
|
230 |
+
"grad_norm": 0.8883898612449065,
|
231 |
+
"learning_rate": 8.345126385981737e-06,
|
232 |
+
"loss": 0.896,
|
233 |
+
"step": 1600
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.5840191133528007,
|
237 |
+
"grad_norm": 0.9045281001313086,
|
238 |
+
"learning_rate": 7.782775833234522e-06,
|
239 |
+
"loss": 0.9033,
|
240 |
+
"step": 1650
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.6017166622422795,
|
244 |
+
"grad_norm": 0.8962005631393778,
|
245 |
+
"learning_rate": 7.227710275938987e-06,
|
246 |
+
"loss": 0.8979,
|
247 |
+
"step": 1700
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.6194142111317582,
|
251 |
+
"grad_norm": 0.8638732399362996,
|
252 |
+
"learning_rate": 6.68175345845119e-06,
|
253 |
+
"loss": 0.8889,
|
254 |
+
"step": 1750
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.637111760021237,
|
258 |
+
"grad_norm": 1.0900143725646285,
|
259 |
+
"learning_rate": 6.146699197107715e-06,
|
260 |
+
"loss": 0.8957,
|
261 |
+
"step": 1800
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.6548093089107159,
|
265 |
+
"grad_norm": 0.9103789814206331,
|
266 |
+
"learning_rate": 5.6243054863949675e-06,
|
267 |
+
"loss": 0.8858,
|
268 |
+
"step": 1850
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 0.6725068578001947,
|
272 |
+
"grad_norm": 0.9300189349082482,
|
273 |
+
"learning_rate": 5.116288722816087e-06,
|
274 |
+
"loss": 0.8822,
|
275 |
+
"step": 1900
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 0.6902044066896735,
|
279 |
+
"grad_norm": 0.8743852772084869,
|
280 |
+
"learning_rate": 4.6243180654337975e-06,
|
281 |
+
"loss": 0.9172,
|
282 |
+
"step": 1950
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 0.7079019555791523,
|
286 |
+
"grad_norm": 0.9546014887497036,
|
287 |
+
"learning_rate": 4.1500099516183555e-06,
|
288 |
+
"loss": 1.1497,
|
289 |
+
"step": 2000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 0.7255995044686311,
|
293 |
+
"grad_norm": 0.9614448138630033,
|
294 |
+
"learning_rate": 3.6949227860198712e-06,
|
295 |
+
"loss": 1.0692,
|
296 |
+
"step": 2050
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.7432970533581099,
|
300 |
+
"grad_norm": 0.848345849582495,
|
301 |
+
"learning_rate": 3.2605518202151577e-06,
|
302 |
+
"loss": 1.0286,
|
303 |
+
"step": 2100
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 0.7609946022475887,
|
307 |
+
"grad_norm": 1.0998684999586237,
|
308 |
+
"learning_rate": 2.8483242398526723e-06,
|
309 |
+
"loss": 0.8803,
|
310 |
+
"step": 2150
|
311 |
+
},
|
312 |
+
{
|
313 |
+
"epoch": 0.7786921511370675,
|
314 |
+
"grad_norm": 0.8279471296348094,
|
315 |
+
"learning_rate": 2.4595944754374723e-06,
|
316 |
+
"loss": 0.8885,
|
317 |
+
"step": 2200
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 0.7963897000265463,
|
321 |
+
"grad_norm": 0.8602056476985009,
|
322 |
+
"learning_rate": 2.0956397521631666e-06,
|
323 |
+
"loss": 0.8704,
|
324 |
+
"step": 2250
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 0.8140872489160251,
|
328 |
+
"grad_norm": 0.9945305020433657,
|
329 |
+
"learning_rate": 1.757655893412622e-06,
|
330 |
+
"loss": 0.8677,
|
331 |
+
"step": 2300
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 0.831784797805504,
|
335 |
+
"grad_norm": 0.850430777727199,
|
336 |
+
"learning_rate": 1.4467533917154842e-06,
|
337 |
+
"loss": 0.877,
|
338 |
+
"step": 2350
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.8494823466949828,
|
342 |
+
"grad_norm": 0.896244091668865,
|
343 |
+
"learning_rate": 1.1639537600719764e-06,
|
344 |
+
"loss": 0.8752,
|
345 |
+
"step": 2400
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 0.8671798955844615,
|
349 |
+
"grad_norm": 0.9590853387394762,
|
350 |
+
"learning_rate": 9.101861756312369e-07,
|
351 |
+
"loss": 0.8737,
|
352 |
+
"step": 2450
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 0.8848774444739403,
|
356 |
+
"grad_norm": 0.805579311652691,
|
357 |
+
"learning_rate": 6.862844267517643e-07,
|
358 |
+
"loss": 0.8792,
|
359 |
+
"step": 2500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 0.9025749933634192,
|
363 |
+
"grad_norm": 0.8767296521208048,
|
364 |
+
"learning_rate": 4.929841734749063e-07,
|
365 |
+
"loss": 0.875,
|
366 |
+
"step": 2550
|
367 |
+
},
|
368 |
+
{
|
369 |
+
"epoch": 0.920272542252898,
|
370 |
+
"grad_norm": 0.9004753221903812,
|
371 |
+
"learning_rate": 3.309205304124552e-07,
|
372 |
+
"loss": 0.8703,
|
373 |
+
"step": 2600
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 0.9379700911423767,
|
377 |
+
"grad_norm": 0.8622876486385493,
|
378 |
+
"learning_rate": 2.0062597999009114e-07,
|
379 |
+
"loss": 0.8815,
|
380 |
+
"step": 2650
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.9556676400318556,
|
384 |
+
"grad_norm": 0.896417353499626,
|
385 |
+
"learning_rate": 1.0252862290301092e-07,
|
386 |
+
"loss": 0.8715,
|
387 |
+
"step": 2700
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.9733651889213344,
|
391 |
+
"grad_norm": 0.9142726428404561,
|
392 |
+
"learning_rate": 3.6950771532126004e-08,
|
393 |
+
"loss": 0.8657,
|
394 |
+
"step": 2750
|
395 |
+
},
|
396 |
+
{
|
397 |
+
"epoch": 0.9910627378108132,
|
398 |
+
"grad_norm": 0.8664983124317569,
|
399 |
+
"learning_rate": 4.1078909423253325e-09,
|
400 |
+
"loss": 0.8791,
|
401 |
+
"step": 2800
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 0.9999115122555526,
|
405 |
+
"step": 2825,
|
406 |
+
"total_flos": 4.909801179132723e+16,
|
407 |
+
"train_loss": 0.9400995777771536,
|
408 |
+
"train_runtime": 88050.3139,
|
409 |
+
"train_samples_per_second": 0.513,
|
410 |
+
"train_steps_per_second": 0.032
|
411 |
+
}
|
412 |
+
],
|
413 |
+
"logging_steps": 50,
|
414 |
+
"max_steps": 2825,
|
415 |
+
"num_input_tokens_seen": 0,
|
416 |
+
"num_train_epochs": 1,
|
417 |
+
"save_steps": 500,
|
418 |
+
"total_flos": 4.909801179132723e+16,
|
419 |
+
"train_batch_size": 4,
|
420 |
+
"trial_name": null,
|
421 |
+
"trial_params": null
|
422 |
+
}
|
llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6c61e1b93e453464c82a151998343e20e07c872d423dcc4a4a2e46bd6e53948b
|
3 |
+
size 7096
|