Jinchen Ge commited on
Commit
7408dc9
1 Parent(s): e8279bb

Add fine-tuned model

Browse files
README.md ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ datasets:
5
+ - Graphcore/gqa-lxmert
6
+ metrics:
7
+ - accuracy
8
+ model-index:
9
+ - name: gqa
10
+ results:
11
+ - task:
12
+ name: Question Answering
13
+ type: question-answering
14
+ dataset:
15
+ name: Graphcore/gqa-lxmert
16
+ type: Graphcore/gqa-lxmert
17
+ args: gqa
18
+ metrics:
19
+ - name: Accuracy
20
+ type: accuracy
21
+ value: 0.5933514030612245
22
+ ---
23
+
24
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
25
+ should probably proofread and complete it, then remove this comment. -->
26
+
27
+ # gqa
28
+
29
+ This model is a fine-tuned version of [unc-nlp/lxmert-base-uncased](https://huggingface.co/unc-nlp/lxmert-base-uncased) on the Graphcore/gqa-lxmert dataset.
30
+ It achieves the following results on the evaluation set:
31
+ - Loss: 1.9326
32
+ - Accuracy: 0.5934
33
+
34
+ ## Model description
35
+
36
+ More information needed
37
+
38
+ ## Intended uses & limitations
39
+
40
+ More information needed
41
+
42
+ ## Training and evaluation data
43
+
44
+ More information needed
45
+
46
+ ## Training procedure
47
+
48
+ ### Training hyperparameters
49
+
50
+ The following hyperparameters were used during training:
51
+ - learning_rate: 1e-05
52
+ - train_batch_size: 1
53
+ - eval_batch_size: 8
54
+ - seed: 42
55
+ - distributed_type: IPU
56
+ - total_train_batch_size: 64
57
+ - total_eval_batch_size: 128
58
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
+ - lr_scheduler_type: linear
60
+ - lr_scheduler_warmup_ratio: 0.1
61
+ - num_epochs: 4.0
62
+ - training precision: Mixed Precision
63
+
64
+ ### Training results
65
+
66
+
67
+
68
+ ### Framework versions
69
+
70
+ - Transformers 4.18.0.dev0
71
+ - Pytorch 1.10.0+cpu
72
+ - Datasets 2.0.0
73
+ - Tokenizers 0.11.6
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "eval_accuracy": 0.5933514030612245,
4
+ "eval_loss": 1.9326171875,
5
+ "eval_runtime": 622.7036,
6
+ "eval_samples": 12576,
7
+ "eval_samples_per_second": 20.196,
8
+ "eval_steps_per_second": 0.159,
9
+ "train_loss": 0.6123406731570221,
10
+ "train_runtime": 29986.2288,
11
+ "train_samples": 943000,
12
+ "train_samples_per_second": 125.791,
13
+ "train_steps_per_second": 1.965
14
+ }
config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "unc-nlp/lxmert-base-uncased",
3
+ "architectures": [
4
+ "PoptorchPipelinedLxmertForQuestionAnswering"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "hidden_act": "gelu",
8
+ "hidden_dropout_prob": 0.1,
9
+ "hidden_size": 768,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 3072,
12
+ "l_layers": 9,
13
+ "layer_norm_eps": 1e-12,
14
+ "max_position_embeddings": 512,
15
+ "model_type": "lxmert",
16
+ "num_attention_heads": 12,
17
+ "num_attr_labels": 400,
18
+ "num_hidden_layers": {
19
+ "cross_encoder": 5,
20
+ "language": 9,
21
+ "vision": 5
22
+ },
23
+ "num_object_labels": 1600,
24
+ "num_qa_labels": 1842,
25
+ "r_layers": 5,
26
+ "task_mask_lm": true,
27
+ "task_matched": true,
28
+ "task_obj_predict": true,
29
+ "task_qa": true,
30
+ "torch_dtype": "float16",
31
+ "transformers_version": "4.18.0.dev0",
32
+ "type_vocab_size": 2,
33
+ "visual_attr_loss": true,
34
+ "visual_feat_dim": 2048,
35
+ "visual_feat_loss": true,
36
+ "visual_loss_normalizer": 6.67,
37
+ "visual_obj_loss": true,
38
+ "visual_pos_dim": 4,
39
+ "vocab_size": 30522,
40
+ "x_layers": 5
41
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "eval_accuracy": 0.5933514030612245,
4
+ "eval_loss": 1.9326171875,
5
+ "eval_runtime": 622.7036,
6
+ "eval_samples": 12576,
7
+ "eval_samples_per_second": 20.196,
8
+ "eval_steps_per_second": 0.159
9
+ }
ipu_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "device_iterations": 2,
3
+ "embedding_serialization_factor": 0,
4
+ "enable_half_first_order_momentum": true,
5
+ "enable_half_partials": true,
6
+ "executable_cache_dir": "./exe_cache",
7
+ "gradient_accumulation_steps": 8,
8
+ "inference_device_iterations": 4,
9
+ "inference_replication_factor": 4,
10
+ "ipus_per_replica": 4,
11
+ "layers_per_ipu": [
12
+ 1
13
+ ],
14
+ "matmul_proportion": [
15
+ 0.25
16
+ ],
17
+ "optimizer_state_offchip": true,
18
+ "optimum_version": "1.0.0",
19
+ "output_mode": "final",
20
+ "profile_dir": "",
21
+ "recompute_checkpoint_every_layer": true,
22
+ "replicated_tensor_sharding": true,
23
+ "replication_factor": 4,
24
+ "seed": 42,
25
+ "use_popdist": false
26
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b88b8a990601f56bb6a472a64e60c9564f27bdd51fb3133fc1bd003bb5bfcf19
3
+ size 424052723
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": "/home/eltoto/unc-nlp/special_tokens_map.json", "full_tokenizer_file": null, "name_or_path": "unc-nlp/lxmert-base-uncased", "tokenizer_class": "LxmertTokenizer"}
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 4.0,
3
+ "train_loss": 0.6123406731570221,
4
+ "train_runtime": 29986.2288,
5
+ "train_samples": 943000,
6
+ "train_samples_per_second": 125.791,
7
+ "train_steps_per_second": 1.965
8
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86f0f5fcb6ddc51360049160f8411161e01b678ea22063511dc8e045e7b61954
3
+ size 2671
vocab.txt ADDED
The diff for this file is too large to render. See raw diff