jackal1586
commited on
Commit
•
9545908
1
Parent(s):
14fb5b4
Add new checkpoints
Browse files- .gitattributes +0 -0
- config.json +28 -0
- flax_model.msgpack +3 -0
- tensorboard/events.out.tfevents.1625801160.t1v-n-a220b410-w-0.244562.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625801750.t1v-n-a220b410-w-0.257638.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625802132.t1v-n-a220b410-w-0.266627.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625802211.t1v-n-a220b410-w-0.268084.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625805762.t1v-n-a220b410-w-0.303880.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625839648.t1v-n-a220b410-w-0.407921.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625840890.t1v-n-a220b410-w-0.422647.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625843300.t1v-n-a220b410-w-0.438643.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625844119.t1v-n-a220b410-w-0.453534.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625845062.t1v-n-a220b410-w-0.467540.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625845925.t1v-n-a220b410-w-0.481564.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625847151.t1v-n-a220b410-w-0.496172.3.v2 +3 -0
- tensorboard/events.out.tfevents.1625855201.t1v-n-a220b410-w-0.521540.3.v2 +3 -0
.gitattributes
CHANGED
File without changes
|
config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "/home/patrick/hugging_face/t5/mt5-base",
|
3 |
+
"architectures": [
|
4 |
+
"T5ForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"d_ff": 2048,
|
7 |
+
"d_kv": 64,
|
8 |
+
"d_model": 768,
|
9 |
+
"decoder_start_token_id": 0,
|
10 |
+
"dropout_rate": 0.1,
|
11 |
+
"eos_token_id": 1,
|
12 |
+
"feed_forward_proj": "gated-gelu",
|
13 |
+
"initializer_factor": 1.0,
|
14 |
+
"is_encoder_decoder": true,
|
15 |
+
"layer_norm_epsilon": 1e-06,
|
16 |
+
"model_type": "mt5",
|
17 |
+
"num_decoder_layers": 12,
|
18 |
+
"num_heads": 12,
|
19 |
+
"num_layers": 12,
|
20 |
+
"output_past": true,
|
21 |
+
"pad_token_id": 0,
|
22 |
+
"relative_attention_num_buckets": 32,
|
23 |
+
"tie_word_embeddings": false,
|
24 |
+
"tokenizer_class": "T5Tokenizer",
|
25 |
+
"transformers_version": "4.9.0.dev0",
|
26 |
+
"use_cache": true,
|
27 |
+
"vocab_size": 28160
|
28 |
+
}
|
flax_model.msgpack
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:87a6643c64e2fb3b149b525d0a4db9c521b6641fdbfa64be12861f3e92c06646
|
3 |
+
size 965944223
|
tensorboard/events.out.tfevents.1625801160.t1v-n-a220b410-w-0.244562.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0411892794c8464e194fe8dcf6f682b522599f837211358ff19b7afda1aefd32
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625801750.t1v-n-a220b410-w-0.257638.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3119cbb31fc1637b238983bef2ca7cb40de86eb02cf5502e84ed191e8d544b2e
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625802132.t1v-n-a220b410-w-0.266627.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a7f05a3a707b81d74ff950595a4c0d223d9c3f636d849a39667fb1ab89b02eea
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625802211.t1v-n-a220b410-w-0.268084.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f735798820597d827c66cf5ec7fff4efe105acea681ac06131240bf9ea0b158d
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625805762.t1v-n-a220b410-w-0.303880.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:92a0f40fb4db5e8ae91e13e0d7fe6c11d0fe3b870c613083f6457a40f3ec7a05
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625839648.t1v-n-a220b410-w-0.407921.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4ab41ddd450a35cb0965b10dfbc4888381449aab588569be0273bc475d1db645
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625840890.t1v-n-a220b410-w-0.422647.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c944177f2c41809358f1655a3ef7f547e2f07ed8df6ad061499eebbb9b021643
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625843300.t1v-n-a220b410-w-0.438643.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:724ff39e23b305b90d50c543528d5bc1ce34a643c068794bcd88e8a34409c037
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625844119.t1v-n-a220b410-w-0.453534.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a138830a224a919271425b919e8469008fc193a6951b1ceaa13144162d5f02d
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625845062.t1v-n-a220b410-w-0.467540.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:af3c5a32a6116f953dcfc91b0de11d98343a5771999222bcb11258664c1a00e9
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625845925.t1v-n-a220b410-w-0.481564.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:06dfaa5e87c361cdf365a03c5bba6bbb7e298d926ee35bd8b3e0035f5290f50d
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625847151.t1v-n-a220b410-w-0.496172.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cf012f0e77c045da96ecb4fe9df2247b94d296dd08f4fc959ed8a56b3e709c09
|
3 |
+
size 40
|
tensorboard/events.out.tfevents.1625855201.t1v-n-a220b410-w-0.521540.3.v2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4cd1aaf39b3e923f725b43dad289743eb791e90d16b611505fcd190efc990a5e
|
3 |
+
size 40
|