Epoch 0 - Val loss 1.1532
Browse files- README.md +12 -0
- config.json +21 -0
README.md
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
pipeline_tag: text-generation
|
4 |
+
tags:
|
5 |
+
- model_hub_mixin
|
6 |
+
- pytorch_model_hub_mixin
|
7 |
+
---
|
8 |
+
|
9 |
+
This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
|
10 |
+
- Code: [More Information Needed]
|
11 |
+
- Paper: [More Information Needed]
|
12 |
+
- Docs: [More Information Needed]
|
config.json
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"att_groups": 4,
|
3 |
+
"att_heads": 8,
|
4 |
+
"att_query_groups": 4,
|
5 |
+
"cross_att_type": "sqa",
|
6 |
+
"embed_dim": 128,
|
7 |
+
"ff_dim": 256,
|
8 |
+
"ff_dropout": 0.1,
|
9 |
+
"init_identity_norm": true,
|
10 |
+
"moe_top_k": 4,
|
11 |
+
"num_experts": 20,
|
12 |
+
"num_layers": 10,
|
13 |
+
"self_att_type": "sqa",
|
14 |
+
"seq_len": 256,
|
15 |
+
"stm_size": 256,
|
16 |
+
"use_flash_attention": false,
|
17 |
+
"use_gated": true,
|
18 |
+
"use_head_norm": true,
|
19 |
+
"use_moe": true,
|
20 |
+
"vocab_size": 7500
|
21 |
+
}
|