Hugging Face's logo

chanyongp
/
colpali_finetuning

Transformers
Safetensors
ColPali
Generated from Trainer
Inference Endpoints
Model card Files Community
chanyongp commited on
Commit
b6d984e
•
1 Parent(s): 6f30c73

End of training

Browse files
Files changed (4) hide show
  1. README.md +173 -0
  2. adapter_config.json +26 -0
  3. adapter_model.safetensors +3 -0
  4. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,173 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: gemma
4
+ base_model: vidore/colpaligemma-3b-pt-448-base
5
+ tags:
6
+ - colpali
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: colpali_finetuning
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # colpali_finetuning
17
+
18
+ This model is a fine-tuned version of [vidore/colpaligemma-3b-pt-448-base](https://huggingface.co/vidore/colpaligemma-3b-pt-448-base) on the vidore/colpali_train_set dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 0.0537
21
+ - Model Preparation Time: 0.0062
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 5e-05
41
+ - train_batch_size: 4
42
+ - eval_batch_size: 4
43
+ - seed: 42
44
+ - gradient_accumulation_steps: 4
45
+ - total_train_batch_size: 16
46
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
+ - lr_scheduler_type: linear
48
+ - lr_scheduler_warmup_steps: 100
49
+ - num_epochs: 1.5
50
+
51
+ ### Training results
52
+
53
+ | Training Loss | Epoch | Step | Validation Loss | Model Preparation Time |
54
+ |:-------------:|:------:|:-----:|:---------------:|:----------------------:|
55
+ | No log | 0.0001 | 1 | 0.0537 | 0.0062 |
56
+ | 0.0236 | 0.0135 | 100 | 0.0564 | 0.0062 |
57
+ | 0.0535 | 0.0271 | 200 | 0.0593 | 0.0062 |
58
+ | 0.0306 | 0.0406 | 300 | 0.0557 | 0.0062 |
59
+ | 0.034 | 0.0541 | 400 | 0.0695 | 0.0062 |
60
+ | 0.0326 | 0.0677 | 500 | 0.0445 | 0.0062 |
61
+ | 0.0465 | 0.0812 | 600 | 0.0527 | 0.0062 |
62
+ | 0.0344 | 0.0948 | 700 | 0.0533 | 0.0062 |
63
+ | 0.0634 | 0.1083 | 800 | 0.0438 | 0.0062 |
64
+ | 0.0474 | 0.1218 | 900 | 0.0455 | 0.0062 |
65
+ | 0.0383 | 0.1354 | 1000 | 0.0490 | 0.0062 |
66
+ | 0.0749 | 0.1489 | 1100 | 0.0460 | 0.0062 |
67
+ | 0.0507 | 0.1624 | 1200 | 0.0361 | 0.0062 |
68
+ | 0.05 | 0.1760 | 1300 | 0.0385 | 0.0062 |
69
+ | 0.0409 | 0.1895 | 1400 | 0.0451 | 0.0062 |
70
+ | 0.037 | 0.2031 | 1500 | 0.0427 | 0.0062 |
71
+ | 0.0593 | 0.2166 | 1600 | 0.0653 | 0.0062 |
72
+ | 0.035 | 0.2301 | 1700 | 0.0596 | 0.0062 |
73
+ | 0.0422 | 0.2437 | 1800 | 0.0471 | 0.0062 |
74
+ | 0.0356 | 0.2572 | 1900 | 0.0498 | 0.0062 |
75
+ | 0.0088 | 0.2707 | 2000 | 0.0608 | 0.0062 |
76
+ | 0.0227 | 0.2843 | 2100 | 0.0720 | 0.0062 |
77
+ | 0.0166 | 0.2978 | 2200 | 0.0734 | 0.0062 |
78
+ | 0.0348 | 0.3113 | 2300 | 0.0589 | 0.0062 |
79
+ | 0.0161 | 0.3249 | 2400 | 0.0596 | 0.0062 |
80
+ | 0.0316 | 0.3384 | 2500 | 0.0499 | 0.0062 |
81
+ | 0.0365 | 0.3520 | 2600 | 0.0543 | 0.0062 |
82
+ | 0.0304 | 0.3655 | 2700 | 0.0556 | 0.0062 |
83
+ | 0.0105 | 0.3790 | 2800 | 0.0545 | 0.0062 |
84
+ | 0.0136 | 0.3926 | 2900 | 0.0545 | 0.0062 |
85
+ | 0.015 | 0.4061 | 3000 | 0.0605 | 0.0062 |
86
+ | 0.0453 | 0.4196 | 3100 | 0.0611 | 0.0062 |
87
+ | 0.0364 | 0.4332 | 3200 | 0.0593 | 0.0062 |
88
+ | 0.0347 | 0.4467 | 3300 | 0.0579 | 0.0062 |
89
+ | 0.0225 | 0.4603 | 3400 | 0.0602 | 0.0062 |
90
+ | 0.0232 | 0.4738 | 3500 | 0.0505 | 0.0062 |
91
+ | 0.0318 | 0.4873 | 3600 | 0.0514 | 0.0062 |
92
+ | 0.0359 | 0.5009 | 3700 | 0.0542 | 0.0062 |
93
+ | 0.0366 | 0.5144 | 3800 | 0.0559 | 0.0062 |
94
+ | 0.0323 | 0.5279 | 3900 | 0.0584 | 0.0062 |
95
+ | 0.0205 | 0.5415 | 4000 | 0.0615 | 0.0062 |
96
+ | 0.0292 | 0.5550 | 4100 | 0.0527 | 0.0062 |
97
+ | 0.0458 | 0.5685 | 4200 | 0.0506 | 0.0062 |
98
+ | 0.0175 | 0.5821 | 4300 | 0.0503 | 0.0062 |
99
+ | 0.0215 | 0.5956 | 4400 | 0.0558 | 0.0062 |
100
+ | 0.0142 | 0.6092 | 4500 | 0.0568 | 0.0062 |
101
+ | 0.0312 | 0.6227 | 4600 | 0.0612 | 0.0062 |
102
+ | 0.0193 | 0.6362 | 4700 | 0.0630 | 0.0062 |
103
+ | 0.0077 | 0.6498 | 4800 | 0.0631 | 0.0062 |
104
+ | 0.0334 | 0.6633 | 4900 | 0.0708 | 0.0062 |
105
+ | 0.0247 | 0.6768 | 5000 | 0.0664 | 0.0062 |
106
+ | 0.0201 | 0.6904 | 5100 | 0.0546 | 0.0062 |
107
+ | 0.0402 | 0.7039 | 5200 | 0.0630 | 0.0062 |
108
+ | 0.0326 | 0.7175 | 5300 | 0.0651 | 0.0062 |
109
+ | 0.0127 | 0.7310 | 5400 | 0.0614 | 0.0062 |
110
+ | 0.0479 | 0.7445 | 5500 | 0.0599 | 0.0062 |
111
+ | 0.0344 | 0.7581 | 5600 | 0.0557 | 0.0062 |
112
+ | 0.016 | 0.7716 | 5700 | 0.0542 | 0.0062 |
113
+ | 0.0194 | 0.7851 | 5800 | 0.0546 | 0.0062 |
114
+ | 0.0236 | 0.7987 | 5900 | 0.0543 | 0.0062 |
115
+ | 0.0285 | 0.8122 | 6000 | 0.0577 | 0.0062 |
116
+ | 0.0128 | 0.8257 | 6100 | 0.0521 | 0.0062 |
117
+ | 0.0218 | 0.8393 | 6200 | 0.0539 | 0.0062 |
118
+ | 0.0501 | 0.8528 | 6300 | 0.0515 | 0.0062 |
119
+ | 0.0456 | 0.8664 | 6400 | 0.0508 | 0.0062 |
120
+ | 0.0247 | 0.8799 | 6500 | 0.0500 | 0.0062 |
121
+ | 0.035 | 0.8934 | 6600 | 0.0516 | 0.0062 |
122
+ | 0.0068 | 0.9070 | 6700 | 0.0502 | 0.0062 |
123
+ | 0.0257 | 0.9205 | 6800 | 0.0527 | 0.0062 |
124
+ | 0.0192 | 0.9340 | 6900 | 0.0513 | 0.0062 |
125
+ | 0.0334 | 0.9476 | 7000 | 0.0551 | 0.0062 |
126
+ | 0.0208 | 0.9611 | 7100 | 0.0544 | 0.0062 |
127
+ | 0.0668 | 0.9747 | 7200 | 0.0510 | 0.0062 |
128
+ | 0.0264 | 0.9882 | 7300 | 0.0481 | 0.0062 |
129
+ | 0.0641 | 1.0017 | 7400 | 0.0486 | 0.0062 |
130
+ | 0.0178 | 1.0153 | 7500 | 0.0473 | 0.0062 |
131
+ | 0.0206 | 1.0288 | 7600 | 0.0502 | 0.0062 |
132
+ | 0.0188 | 1.0423 | 7700 | 0.0537 | 0.0062 |
133
+ | 0.0378 | 1.0559 | 7800 | 0.0502 | 0.0062 |
134
+ | 0.0313 | 1.0694 | 7900 | 0.0571 | 0.0062 |
135
+ | 0.0169 | 1.0829 | 8000 | 0.0586 | 0.0062 |
136
+ | 0.0164 | 1.0965 | 8100 | 0.0580 | 0.0062 |
137
+ | 0.0327 | 1.1100 | 8200 | 0.0540 | 0.0062 |
138
+ | 0.0153 | 1.1236 | 8300 | 0.0507 | 0.0062 |
139
+ | 0.0305 | 1.1371 | 8400 | 0.0542 | 0.0062 |
140
+ | 0.0279 | 1.1506 | 8500 | 0.0532 | 0.0062 |
141
+ | 0.0081 | 1.1642 | 8600 | 0.0552 | 0.0062 |
142
+ | 0.027 | 1.1777 | 8700 | 0.0545 | 0.0062 |
143
+ | 0.0112 | 1.1912 | 8800 | 0.0551 | 0.0062 |
144
+ | 0.0312 | 1.2048 | 8900 | 0.0564 | 0.0062 |
145
+ | 0.0244 | 1.2183 | 9000 | 0.0538 | 0.0062 |
146
+ | 0.0274 | 1.2319 | 9100 | 0.0529 | 0.0062 |
147
+ | 0.0351 | 1.2454 | 9200 | 0.0531 | 0.0062 |
148
+ | 0.0172 | 1.2589 | 9300 | 0.0532 | 0.0062 |
149
+ | 0.005 | 1.2725 | 9400 | 0.0513 | 0.0062 |
150
+ | 0.0195 | 1.2860 | 9500 | 0.0549 | 0.0062 |
151
+ | 0.0062 | 1.2995 | 9600 | 0.0558 | 0.0062 |
152
+ | 0.0234 | 1.3131 | 9700 | 0.0558 | 0.0062 |
153
+ | 0.0157 | 1.3266 | 9800 | 0.0564 | 0.0062 |
154
+ | 0.0248 | 1.3401 | 9900 | 0.0556 | 0.0062 |
155
+ | 0.0098 | 1.3537 | 10000 | 0.0536 | 0.0062 |
156
+ | 0.0133 | 1.3672 | 10100 | 0.0525 | 0.0062 |
157
+ | 0.0187 | 1.3808 | 10200 | 0.0526 | 0.0062 |
158
+ | 0.0088 | 1.3943 | 10300 | 0.0509 | 0.0062 |
159
+ | 0.0315 | 1.4078 | 10400 | 0.0534 | 0.0062 |
160
+ | 0.0219 | 1.4214 | 10500 | 0.0537 | 0.0062 |
161
+ | 0.0225 | 1.4349 | 10600 | 0.0548 | 0.0062 |
162
+ | 0.0338 | 1.4484 | 10700 | 0.0545 | 0.0062 |
163
+ | 0.029 | 1.4620 | 10800 | 0.0539 | 0.0062 |
164
+ | 0.0354 | 1.4755 | 10900 | 0.0537 | 0.0062 |
165
+ | 0.0214 | 1.4891 | 11000 | 0.0536 | 0.0062 |
166
+
167
+
168
+ ### Framework versions
169
+
170
+ - Transformers 4.45.2
171
+ - Pytorch 2.4.1+cu121
172
+ - Datasets 3.0.2
173
+ - Tokenizers 0.20.1
adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "vidore/colpaligemma-3b-pt-448-base",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": "gaussian",
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 32,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "(.*(language_model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)",
23
+ "task_type": "FEATURE_EXTRACTION",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e14542573d8691f464f1f5fd132a9072ae0885d325c3a9e114d79b4ed2e9050
3
+ size 157071680
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78e304f57048465cd51afd3d526492bd495b038d6639c08c600465941297e120
3
+ size 5240