move model files to root
Browse files- README.md +2 -5
- config.json +127 -0
- flax_model.msgpack +3 -0
- output/single_latent +1 -0
- setup_venv.sh → setup_tpu_vm_venv.sh +0 -0
README.md
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
# T5-VAE-Python (flax)
|
2 |
|
3 |
A Transformer-VAE made using flax.
|
4 |
|
@@ -10,7 +10,4 @@ Builds on T5, using an autoencoder to convert it into an MMD-VAE.
|
|
10 |
|
11 |
## Setup
|
12 |
|
13 |
-
|
14 |
-
|
15 |
-
- [ ] Find dataset storage site.
|
16 |
-
- [ ] Ask JAX team for dataset storage.
|
|
|
1 |
+
# T5-VAE-Python (flax)
|
2 |
|
3 |
A Transformer-VAE made using flax.
|
4 |
|
|
|
10 |
|
11 |
## Setup
|
12 |
|
13 |
+
Run `setup_tpu_vm_venv.sh` to setup a virtual enviroment.
|
|
|
|
|
|
config.json
ADDED
@@ -0,0 +1,127 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_special_tokens": false,
|
3 |
+
"architectures": [
|
4 |
+
"T5VaeForAutoencoding"
|
5 |
+
],
|
6 |
+
"config_path": null,
|
7 |
+
"decoder_start_token_id": 0,
|
8 |
+
"dtype": "float32",
|
9 |
+
"eos_token_id": 1,
|
10 |
+
"latent_token_size": 32,
|
11 |
+
"model_name_or_path": null,
|
12 |
+
"model_type": "transformer_vae",
|
13 |
+
"n_latent_tokens": 1,
|
14 |
+
"pad_token_id": 0,
|
15 |
+
"set_seq_size": 60,
|
16 |
+
"t5": {
|
17 |
+
"_name_or_path": "",
|
18 |
+
"add_cross_attention": false,
|
19 |
+
"architectures": [
|
20 |
+
"T5WithLMHeadModel"
|
21 |
+
],
|
22 |
+
"bad_words_ids": null,
|
23 |
+
"bos_token_id": null,
|
24 |
+
"chunk_size_feed_forward": 0,
|
25 |
+
"d_ff": 3072,
|
26 |
+
"d_kv": 64,
|
27 |
+
"d_model": 768,
|
28 |
+
"decoder_start_token_id": 0,
|
29 |
+
"diversity_penalty": 0.0,
|
30 |
+
"do_sample": false,
|
31 |
+
"dropout_rate": 0.1,
|
32 |
+
"early_stopping": false,
|
33 |
+
"encoder_no_repeat_ngram_size": 0,
|
34 |
+
"eos_token_id": 1,
|
35 |
+
"feed_forward_proj": "relu",
|
36 |
+
"finetuning_task": null,
|
37 |
+
"forced_bos_token_id": null,
|
38 |
+
"forced_eos_token_id": null,
|
39 |
+
"gradient_checkpointing": false,
|
40 |
+
"id2label": {
|
41 |
+
"0": "LABEL_0",
|
42 |
+
"1": "LABEL_1"
|
43 |
+
},
|
44 |
+
"initializer_factor": 1.0,
|
45 |
+
"is_decoder": false,
|
46 |
+
"is_encoder_decoder": true,
|
47 |
+
"label2id": {
|
48 |
+
"LABEL_0": 0,
|
49 |
+
"LABEL_1": 1
|
50 |
+
},
|
51 |
+
"layer_norm_epsilon": 1e-06,
|
52 |
+
"length_penalty": 1.0,
|
53 |
+
"max_length": 20,
|
54 |
+
"min_length": 0,
|
55 |
+
"model_type": "t5",
|
56 |
+
"n_positions": 512,
|
57 |
+
"no_repeat_ngram_size": 0,
|
58 |
+
"num_beam_groups": 1,
|
59 |
+
"num_beams": 1,
|
60 |
+
"num_decoder_layers": 12,
|
61 |
+
"num_heads": 12,
|
62 |
+
"num_layers": 12,
|
63 |
+
"num_return_sequences": 1,
|
64 |
+
"output_attentions": false,
|
65 |
+
"output_hidden_states": false,
|
66 |
+
"output_past": true,
|
67 |
+
"output_scores": false,
|
68 |
+
"pad_token_id": 0,
|
69 |
+
"prefix": null,
|
70 |
+
"problem_type": null,
|
71 |
+
"pruned_heads": {},
|
72 |
+
"relative_attention_num_buckets": 32,
|
73 |
+
"remove_invalid_values": false,
|
74 |
+
"repetition_penalty": 1.0,
|
75 |
+
"return_dict": true,
|
76 |
+
"return_dict_in_generate": false,
|
77 |
+
"sep_token_id": null,
|
78 |
+
"task_specific_params": {
|
79 |
+
"summarization": {
|
80 |
+
"early_stopping": true,
|
81 |
+
"length_penalty": 2.0,
|
82 |
+
"max_length": 200,
|
83 |
+
"min_length": 30,
|
84 |
+
"no_repeat_ngram_size": 3,
|
85 |
+
"num_beams": 4,
|
86 |
+
"prefix": "summarize: "
|
87 |
+
},
|
88 |
+
"translation_en_to_de": {
|
89 |
+
"early_stopping": true,
|
90 |
+
"max_length": 300,
|
91 |
+
"num_beams": 4,
|
92 |
+
"prefix": "translate English to German: "
|
93 |
+
},
|
94 |
+
"translation_en_to_fr": {
|
95 |
+
"early_stopping": true,
|
96 |
+
"max_length": 300,
|
97 |
+
"num_beams": 4,
|
98 |
+
"prefix": "translate English to French: "
|
99 |
+
},
|
100 |
+
"translation_en_to_ro": {
|
101 |
+
"early_stopping": true,
|
102 |
+
"max_length": 300,
|
103 |
+
"num_beams": 4,
|
104 |
+
"prefix": "translate English to Romanian: "
|
105 |
+
}
|
106 |
+
},
|
107 |
+
"temperature": 1.0,
|
108 |
+
"tie_encoder_decoder": false,
|
109 |
+
"tie_word_embeddings": true,
|
110 |
+
"tokenizer_class": null,
|
111 |
+
"top_k": 50,
|
112 |
+
"top_p": 1.0,
|
113 |
+
"torch_dtype": null,
|
114 |
+
"torchscript": false,
|
115 |
+
"transformers_version": "4.9.0.dev0",
|
116 |
+
"use_bfloat16": false,
|
117 |
+
"use_cache": true,
|
118 |
+
"vocab_size": 32100
|
119 |
+
},
|
120 |
+
"tokenizer_name": null,
|
121 |
+
"transformers_version": null,
|
122 |
+
"use_cache": true,
|
123 |
+
"use_fast_tokenizer": true,
|
124 |
+
"vae_decoder_model": "",
|
125 |
+
"vae_encoder_model": "",
|
126 |
+
"vocab_size": 32100
|
127 |
+
}
|
flax_model.msgpack
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:78df6b3743922890773961fedfcd96f0c6af3fb53a81dd02b2ac50da99424ee8
|
3 |
+
size 891745503
|
output/single_latent
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
Subproject commit 006e94e4a97d7c716219d69af790d7bd8a7db319
|
setup_venv.sh → setup_tpu_vm_venv.sh
RENAMED
File without changes
|