Sync execution step: 164000
Browse files- checkpoint/config.json +2 -3
- checkpoint/generation_config.json +1 -1
- checkpoint/model.safetensors +2 -2
- checkpoint/optimizer.bin +2 -2
- checkpoint/scheduler.bin +1 -1
- checkpoint/tokenizer_config.json +0 -1
- checkpoint/train.json +2 -2
checkpoint/config.json
CHANGED
|
@@ -8,7 +8,7 @@
|
|
| 8 |
"attn_scale": 0.1,
|
| 9 |
"attn_temperature_tuning": true,
|
| 10 |
"bos_token_id": 1,
|
| 11 |
-
"dtype": "
|
| 12 |
"eos_token_id": 2,
|
| 13 |
"floor_scale": 8192,
|
| 14 |
"head_dim": 128,
|
|
@@ -56,7 +56,6 @@
|
|
| 56 |
14,
|
| 57 |
15
|
| 58 |
],
|
| 59 |
-
"no_rope_layer_interval": 4,
|
| 60 |
"no_rope_layers": [
|
| 61 |
1,
|
| 62 |
1,
|
|
@@ -123,7 +122,7 @@
|
|
| 123 |
"router_jitter_noise": 0.0,
|
| 124 |
"sliding_window": 512,
|
| 125 |
"tie_word_embeddings": false,
|
| 126 |
-
"transformers_version": "5.
|
| 127 |
"use_cache": false,
|
| 128 |
"use_qk_norm": true,
|
| 129 |
"vocab_size": 202048
|
|
|
|
| 8 |
"attn_scale": 0.1,
|
| 9 |
"attn_temperature_tuning": true,
|
| 10 |
"bos_token_id": 1,
|
| 11 |
+
"dtype": "float32",
|
| 12 |
"eos_token_id": 2,
|
| 13 |
"floor_scale": 8192,
|
| 14 |
"head_dim": 128,
|
|
|
|
| 56 |
14,
|
| 57 |
15
|
| 58 |
],
|
|
|
|
| 59 |
"no_rope_layers": [
|
| 60 |
1,
|
| 61 |
1,
|
|
|
|
| 122 |
"router_jitter_noise": 0.0,
|
| 123 |
"sliding_window": 512,
|
| 124 |
"tie_word_embeddings": false,
|
| 125 |
+
"transformers_version": "5.0.0",
|
| 126 |
"use_cache": false,
|
| 127 |
"use_qk_norm": true,
|
| 128 |
"vocab_size": 202048
|
checkpoint/generation_config.json
CHANGED
|
@@ -2,5 +2,5 @@
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
| 5 |
-
"transformers_version": "5.
|
| 6 |
}
|
|
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "5.0.0"
|
| 6 |
}
|
checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d801a2b67a3999d68d5ff4b73eb851dc54c32e50fcbe411995e5ca3307689a7e
|
| 3 |
+
size 954310784
|
checkpoint/optimizer.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e7c6852e25749235da6c65e57a5b404e5f9cedba01501092fa3306cc4f974ed9
|
| 3 |
+
size 1908743179
|
checkpoint/scheduler.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:29117a4a7ed67f172c076bb96577328cb108ea2954bacfc1660fc36ae475e8f4
|
| 3 |
size 1465
|
checkpoint/tokenizer_config.json
CHANGED
|
@@ -4,7 +4,6 @@
|
|
| 4 |
"clean_up_tokenization_spaces": false,
|
| 5 |
"eos_token": "<|eot|>",
|
| 6 |
"is_local": false,
|
| 7 |
-
"local_files_only": false,
|
| 8 |
"max_length": 1024,
|
| 9 |
"model_input_names": [
|
| 10 |
"input_ids",
|
|
|
|
| 4 |
"clean_up_tokenization_spaces": false,
|
| 5 |
"eos_token": "<|eot|>",
|
| 6 |
"is_local": false,
|
|
|
|
| 7 |
"max_length": 1024,
|
| 8 |
"model_input_names": [
|
| 9 |
"input_ids",
|
checkpoint/train.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"global_step":
|
| 3 |
"model_name_or_path": "GTKING/ZFusionAI_Base",
|
| 4 |
-
"saved_at_time": "2026-06-18
|
| 5 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"global_step": 164000,
|
| 3 |
"model_name_or_path": "GTKING/ZFusionAI_Base",
|
| 4 |
+
"saved_at_time": "2026-06-18 19:52:47"
|
| 5 |
}
|