Training in progress, step 25000
Browse files- config.json +29 -0
- model.safetensors +3 -0
- runs/Feb19_13-17-18_kiran-ku/events.out.tfevents.1739971038.kiran-ku +3 -0
- runs/Feb19_13-17-35_kiran-ku/events.out.tfevents.1739971059.kiran-ku +3 -0
- runs/Feb19_13-19-44_kiran-ku/events.out.tfevents.1739971188.kiran-ku +3 -0
- runs/Feb19_13-24-07_kiran-ku/events.out.tfevents.1739971448.kiran-ku +3 -0
- runs/Feb19_13-24-53_kiran-ku/events.out.tfevents.1739971496.kiran-ku +3 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"T5ForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"classifier_dropout": 0.0,
|
| 6 |
+
"d_ff": 1024,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 256,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dense_act_fn": "relu",
|
| 11 |
+
"dropout_rate": 0.1,
|
| 12 |
+
"eos_token_id": 1,
|
| 13 |
+
"feed_forward_proj": "relu",
|
| 14 |
+
"initializer_factor": 1.0,
|
| 15 |
+
"is_encoder_decoder": true,
|
| 16 |
+
"is_gated_act": false,
|
| 17 |
+
"layer_norm_epsilon": 1e-06,
|
| 18 |
+
"model_type": "t5",
|
| 19 |
+
"num_decoder_layers": 4,
|
| 20 |
+
"num_heads": 4,
|
| 21 |
+
"num_layers": 4,
|
| 22 |
+
"pad_token_id": 0,
|
| 23 |
+
"relative_attention_max_distance": 128,
|
| 24 |
+
"relative_attention_num_buckets": 32,
|
| 25 |
+
"torch_dtype": "float32",
|
| 26 |
+
"transformers_version": "4.50.0.dev0",
|
| 27 |
+
"use_cache": true,
|
| 28 |
+
"vocab_size": 96
|
| 29 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b96c757b2cd4716ab973ed5de60912b0c240b38b5501e24fd2d0b590f84035db
|
| 3 |
+
size 29492312
|
runs/Feb19_13-17-18_kiran-ku/events.out.tfevents.1739971038.kiran-ku
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db8edf6d0fad9b1c6e45f139934960fb43c4128b1713e81577166c6813d16e7c
|
| 3 |
+
size 4963
|
runs/Feb19_13-17-35_kiran-ku/events.out.tfevents.1739971059.kiran-ku
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e5d606f79815c18e56177a79bdcb8ee1407af4121f31de41aeca56f9f8bb8dc
|
| 3 |
+
size 4963
|
runs/Feb19_13-19-44_kiran-ku/events.out.tfevents.1739971188.kiran-ku
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:23cad726f4abe6ca987886b079133f5fc3df9f94ec6f1bbdcea959d4ff74db57
|
| 3 |
+
size 4960
|
runs/Feb19_13-24-07_kiran-ku/events.out.tfevents.1739971448.kiran-ku
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7aa6e9558a88d728ecff6e832cd6cb515d18c5cb8fb91bbc806785b459ee659d
|
| 3 |
+
size 4960
|
runs/Feb19_13-24-53_kiran-ku/events.out.tfevents.1739971496.kiran-ku
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5422ce1a4f3276710cd62ffe77061f9e72d9dadb6436f509b36b588ec5b379c3
|
| 3 |
+
size 15015
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b4aac925e98b152c900bec06c5c0230d48d69a1962c60f536bcb31a0b13f5b74
|
| 3 |
+
size 5304
|