Upload folder using huggingface_hub
Browse files- checkpoint-5400/config.json +16 -0
- checkpoint-5400/generation_config.json +6 -0
- checkpoint-5400/model.safetensors +3 -0
- checkpoint-5400/optimizer.pt +3 -0
- checkpoint-5400/rng_state.pth +3 -0
- checkpoint-5400/scaler.pt +3 -0
- checkpoint-5400/scheduler.pt +3 -0
- checkpoint-5400/trainer_state.json +0 -0
- checkpoint-5400/training_args.bin +3 -0
checkpoint-5400/config.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"FastHyperNet125mForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"dtype": "float32",
|
| 6 |
+
"hidden_size": 768,
|
| 7 |
+
"hyperbolic_dim": 64,
|
| 8 |
+
"intermediate_size": 2048,
|
| 9 |
+
"max_position_embeddings": 512,
|
| 10 |
+
"model_type": "fasthypernet_125m_tagless",
|
| 11 |
+
"num_attention_heads": 12,
|
| 12 |
+
"num_hidden_layers": 16,
|
| 13 |
+
"transformers_version": "5.13.1",
|
| 14 |
+
"use_cache": false,
|
| 15 |
+
"vocab_size": 8192
|
| 16 |
+
}
|
checkpoint-5400/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"output_attentions": false,
|
| 4 |
+
"output_hidden_states": false,
|
| 5 |
+
"transformers_version": "5.13.1"
|
| 6 |
+
}
|
checkpoint-5400/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e7d1c8cbf80e42f7b880f377b7173cda5c14f13072bdbee53b6697f7fca90c99
|
| 3 |
+
size 542849784
|
checkpoint-5400/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:72d338d7d558a4d40455c0feb3350e600bbe44df0c4e7d7fd5e15db89f95ebc1
|
| 3 |
+
size 273770959
|
checkpoint-5400/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f066b6810f65c3da5de1924550d429ff20ad03263bb1b73200643b753dc0d30c
|
| 3 |
+
size 14645
|
checkpoint-5400/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f149241b7033e2875564f717ff532c3642c294c30df6e11154cf6ff6a1fef243
|
| 3 |
+
size 1383
|
checkpoint-5400/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7bc2757e68d2eb2d763baf03ff77a2d23d0a2a1b88e62cced3f6e602b09f462f
|
| 3 |
+
size 1465
|
checkpoint-5400/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-5400/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cbd241599217f087cf90c5afde8d5857bb2c3234f97a197910829d91c3d8c1f7
|
| 3 |
+
size 5137
|