Edward Gow-Smith
commited on
Commit
•
ede4aa7
1
Parent(s):
7df97a9
added model
Browse files- all_results.json +24 -0
- config.json +42 -0
- eval_results.json +13 -0
- pytorch_model.bin +3 -0
- results.txt +1 -0
- special_tokens_map.json +1 -0
- spiece.model +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- train_results.json +14 -0
- training_args.bin +3 -0
all_results.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_accuracy": 0.7918455004692078,
|
4 |
+
"eval_f1": 0.8264758497316637,
|
5 |
+
"eval_loss": 1.1171678304672241,
|
6 |
+
"eval_mem_cpu_alloc_delta": 4096,
|
7 |
+
"eval_mem_cpu_peaked_delta": 0,
|
8 |
+
"eval_mem_gpu_alloc_delta": 0,
|
9 |
+
"eval_mem_gpu_peaked_delta": 97849856,
|
10 |
+
"eval_runtime": 9.7537,
|
11 |
+
"eval_samples": 466,
|
12 |
+
"eval_samples_per_second": 47.777,
|
13 |
+
"init_mem_cpu_alloc_delta": 894615552,
|
14 |
+
"init_mem_cpu_peaked_delta": 464302080,
|
15 |
+
"init_mem_gpu_alloc_delta": 469504512,
|
16 |
+
"init_mem_gpu_peaked_delta": 0,
|
17 |
+
"train_mem_cpu_alloc_delta": 32620544,
|
18 |
+
"train_mem_cpu_peaked_delta": 0,
|
19 |
+
"train_mem_gpu_alloc_delta": 1454737920,
|
20 |
+
"train_mem_gpu_peaked_delta": 4411375616,
|
21 |
+
"train_runtime": 539.9891,
|
22 |
+
"train_samples": 3327,
|
23 |
+
"train_samples_per_second": 0.578
|
24 |
+
}
|
config.json
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "xlnet-base-cased",
|
3 |
+
"architectures": [
|
4 |
+
"XLNetForSequenceClassification"
|
5 |
+
],
|
6 |
+
"attn_type": "bi",
|
7 |
+
"bi_data": false,
|
8 |
+
"bos_token_id": 1,
|
9 |
+
"clamp_len": -1,
|
10 |
+
"d_head": 64,
|
11 |
+
"d_inner": 3072,
|
12 |
+
"d_model": 768,
|
13 |
+
"dropout": 0.1,
|
14 |
+
"end_n_top": 5,
|
15 |
+
"eos_token_id": 2,
|
16 |
+
"ff_activation": "gelu",
|
17 |
+
"initializer_range": 0.02,
|
18 |
+
"layer_norm_eps": 1e-12,
|
19 |
+
"mem_len": null,
|
20 |
+
"model_type": "xlnet",
|
21 |
+
"n_head": 12,
|
22 |
+
"n_layer": 12,
|
23 |
+
"pad_token_id": 5,
|
24 |
+
"reuse_len": null,
|
25 |
+
"same_length": false,
|
26 |
+
"start_n_top": 5,
|
27 |
+
"summary_activation": "tanh",
|
28 |
+
"summary_last_dropout": 0.1,
|
29 |
+
"summary_type": "last",
|
30 |
+
"summary_use_proj": true,
|
31 |
+
"task_specific_params": {
|
32 |
+
"text-generation": {
|
33 |
+
"do_sample": true,
|
34 |
+
"max_length": 250
|
35 |
+
}
|
36 |
+
},
|
37 |
+
"transformers_version": "4.6.0.dev0",
|
38 |
+
"untie_r": true,
|
39 |
+
"use_mems_eval": true,
|
40 |
+
"use_mems_train": false,
|
41 |
+
"vocab_size": 32000
|
42 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_accuracy": 0.7918455004692078,
|
4 |
+
"eval_f1": 0.8264758497316637,
|
5 |
+
"eval_loss": 1.1171678304672241,
|
6 |
+
"eval_mem_cpu_alloc_delta": 4096,
|
7 |
+
"eval_mem_cpu_peaked_delta": 0,
|
8 |
+
"eval_mem_gpu_alloc_delta": 0,
|
9 |
+
"eval_mem_gpu_peaked_delta": 97849856,
|
10 |
+
"eval_runtime": 9.7537,
|
11 |
+
"eval_samples": 466,
|
12 |
+
"eval_samples_per_second": 47.777
|
13 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:176933adce9648ad59d1a362cdaa50d88b5f82db614637bad4c88b52ba905a7b
|
3 |
+
size 469322193
|
results.txt
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"results": [{"epoch": 3.0, "eval_accuracy": 0.7918455004692078, "eval_f1": 0.8264758497316637}], "seed": 1, "train_path": "EMNLP-2021/data/v3/SentenceClassificationData/FalseTrue-0/train.csv", "dev_path": "EMNLP-2021/data/v3/SentenceClassificationData/FalseTrue-0/dev.csv", "model_name": "xlnet-base-cased"}
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "<sep>", "pad_token": "<pad>", "cls_token": "<cls>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}, "additional_special_tokens": ["<eop>", "<eod>"]}
|
spiece.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1f8c1c0bc2854d1af911a8550288c1258af5ba50277f3a5c829b98eb86fc5646
|
3 |
+
size 798011
|
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": false, "remove_space": true, "keep_accents": false, "bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "<sep>", "pad_token": "<pad>", "cls_token": "<cls>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "additional_special_tokens": ["<eop>", "<eod>"], "special_tokens_map_file": null, "name_or_path": "xlnet-base-cased"}
|
train_results.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"init_mem_cpu_alloc_delta": 894615552,
|
4 |
+
"init_mem_cpu_peaked_delta": 464302080,
|
5 |
+
"init_mem_gpu_alloc_delta": 469504512,
|
6 |
+
"init_mem_gpu_peaked_delta": 0,
|
7 |
+
"train_mem_cpu_alloc_delta": 32620544,
|
8 |
+
"train_mem_cpu_peaked_delta": 0,
|
9 |
+
"train_mem_gpu_alloc_delta": 1454737920,
|
10 |
+
"train_mem_gpu_peaked_delta": 4411375616,
|
11 |
+
"train_runtime": 539.9891,
|
12 |
+
"train_samples": 3327,
|
13 |
+
"train_samples_per_second": 0.578
|
14 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f3c09914f464a5c17a29881937f37fc0082cef7defa0e1bd0d21a8dfbafabb28
|
3 |
+
size 2415
|