Shashank Sonkar
commited on
Commit
•
9a1b730
1
Parent(s):
7f69358
add model
Browse files- config.json +52 -0
- pytorch_model.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "transfo-xl-wt103",
|
| 3 |
+
"adaptive": true,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"HigherPower"
|
| 6 |
+
],
|
| 7 |
+
"attn_type": 0,
|
| 8 |
+
"clamp_len": 1000,
|
| 9 |
+
"cutoffs": [
|
| 10 |
+
20000,
|
| 11 |
+
40000,
|
| 12 |
+
200000
|
| 13 |
+
],
|
| 14 |
+
"d_embed": 1024,
|
| 15 |
+
"d_head": 64,
|
| 16 |
+
"d_inner": 4096,
|
| 17 |
+
"d_model": 1024,
|
| 18 |
+
"div_val": 4,
|
| 19 |
+
"dropatt": 0.0,
|
| 20 |
+
"dropout": 0.1,
|
| 21 |
+
"eos_token_id": 0,
|
| 22 |
+
"ext_len": 0,
|
| 23 |
+
"init": "normal",
|
| 24 |
+
"init_range": 0.01,
|
| 25 |
+
"init_std": 0.02,
|
| 26 |
+
"layer_norm_epsilon": 1e-05,
|
| 27 |
+
"mem_len": 1600,
|
| 28 |
+
"model_type": "transfo-xl",
|
| 29 |
+
"n_head": 16,
|
| 30 |
+
"n_layer": 18,
|
| 31 |
+
"pre_lnorm": false,
|
| 32 |
+
"proj_init_std": 0.01,
|
| 33 |
+
"same_length": true,
|
| 34 |
+
"sample_softmax": -1,
|
| 35 |
+
"task_specific_params": {
|
| 36 |
+
"text-generation": {
|
| 37 |
+
"do_sample": true,
|
| 38 |
+
"max_length": 250
|
| 39 |
+
}
|
| 40 |
+
},
|
| 41 |
+
"tgt_len": 128,
|
| 42 |
+
"tie_projs": [
|
| 43 |
+
false,
|
| 44 |
+
true,
|
| 45 |
+
true,
|
| 46 |
+
true
|
| 47 |
+
],
|
| 48 |
+
"tie_weight": true,
|
| 49 |
+
"transformers_version": "4.6.1",
|
| 50 |
+
"untie_r": true,
|
| 51 |
+
"vocab_size": 267735
|
| 52 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0260b43635980029dfc6581eefa35038517ed7a0e3a531db9c5e11534fade5af
|
| 3 |
+
size 1355119499
|