robingeibel
commited on
Commit
•
9a900fd
1
Parent(s):
ce3937c
End of training
Browse files
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/reformer-crime-and-punishment",
|
3 |
+
"architectures": [
|
4 |
+
"ReformerForMaskedLM"
|
5 |
+
],
|
6 |
+
"attention_head_size": 64,
|
7 |
+
"attention_probs_dropout_prob": 0.1,
|
8 |
+
"attn_layers": [
|
9 |
+
"local",
|
10 |
+
"lsh",
|
11 |
+
"local",
|
12 |
+
"lsh",
|
13 |
+
"local",
|
14 |
+
"lsh"
|
15 |
+
],
|
16 |
+
"axial_norm_std": 1.0,
|
17 |
+
"axial_pos_embds": false,
|
18 |
+
"axial_pos_embds_dim": [
|
19 |
+
128,
|
20 |
+
128
|
21 |
+
],
|
22 |
+
"axial_pos_shape": [
|
23 |
+
512,
|
24 |
+
1024
|
25 |
+
],
|
26 |
+
"chunk_size_lm_head": 0,
|
27 |
+
"classifier_dropout": null,
|
28 |
+
"eos_token_id": 2,
|
29 |
+
"feed_forward_size": 512,
|
30 |
+
"hash_seed": null,
|
31 |
+
"hidden_act": "relu",
|
32 |
+
"hidden_dropout_prob": 0.05,
|
33 |
+
"hidden_size": 256,
|
34 |
+
"initializer_range": 0.02,
|
35 |
+
"intermediate_size": 3072,
|
36 |
+
"layer_norm_eps": 1e-12,
|
37 |
+
"local_attention_probs_dropout_prob": 0.05,
|
38 |
+
"local_attn_chunk_length": 64,
|
39 |
+
"local_num_chunks_after": 0,
|
40 |
+
"local_num_chunks_before": 1,
|
41 |
+
"lsh_attention_probs_dropout_prob": 0.0,
|
42 |
+
"lsh_attn_chunk_length": 64,
|
43 |
+
"lsh_num_chunks_after": 0,
|
44 |
+
"lsh_num_chunks_before": 1,
|
45 |
+
"max_position_embeddings": 16384,
|
46 |
+
"model_type": "reformer",
|
47 |
+
"num_attention_heads": 2,
|
48 |
+
"num_buckets": [
|
49 |
+
64,
|
50 |
+
128
|
51 |
+
],
|
52 |
+
"num_chunks_after": 0,
|
53 |
+
"num_chunks_before": 1,
|
54 |
+
"num_hashes": 1,
|
55 |
+
"num_hidden_layers": 6,
|
56 |
+
"output_past": true,
|
57 |
+
"pad_token_id": 0,
|
58 |
+
"task_specific_params": {
|
59 |
+
"text-generation": {
|
60 |
+
"do_sample": true,
|
61 |
+
"max_length": 100
|
62 |
+
}
|
63 |
+
},
|
64 |
+
"tie_word_embeddings": false,
|
65 |
+
"torch_dtype": "float32",
|
66 |
+
"transformers_version": "4.20.1",
|
67 |
+
"use_cache": true,
|
68 |
+
"vocab_size": 320
|
69 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a05b09e31fd437776b4039cfeeae0a4e70ba7fa18558c5205f39d64d3d1c8275
|
3 |
+
size 26883982
|
runs/Jul12_09-45-21_9e1a41ae41c3/1657619140.4909124/events.out.tfevents.1657619140.9e1a41ae41c3.96.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b095b4408c1ed8e32679b8d91abb8488c2be4bfbea5b5739c4c3b208502a1039
|
3 |
+
size 5375
|
runs/Jul12_09-45-21_9e1a41ae41c3/events.out.tfevents.1657619140.9e1a41ae41c3.96.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1dc62057c3f038bf4e78c73b092fe756a219fb6cc7948fe23836cfdfbcb1f123
|
3 |
+
size 14763
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ec7f58fa8b948d5ff6cbf0d68ad8f9de395ed0cf0d3925e5089c005dba729c71
|
3 |
+
size 3375
|