Update training_config.json
Browse files- training_config.json +1 -28
training_config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"model_name_or_path": "
|
| 3 |
"output_dir": "./checkpoints/helion-2.5-rnd",
|
| 4 |
"overwrite_output_dir": true,
|
| 5 |
"do_train": true,
|
|
@@ -102,33 +102,6 @@
|
|
| 102 |
"ddp_find_unused_parameters": false,
|
| 103 |
"ddp_bucket_cap_mb": null,
|
| 104 |
"ddp_broadcast_buffers": null,
|
| 105 |
-
"dataset_config": {
|
| 106 |
-
"scientific_papers": {
|
| 107 |
-
"path": "arxiv_papers",
|
| 108 |
-
"split": "train",
|
| 109 |
-
"weight": 0.25
|
| 110 |
-
},
|
| 111 |
-
"code_repositories": {
|
| 112 |
-
"path": "the-stack-dedup",
|
| 113 |
-
"split": "train",
|
| 114 |
-
"weight": 0.25
|
| 115 |
-
},
|
| 116 |
-
"mathematical_proofs": {
|
| 117 |
-
"path": "math_qa",
|
| 118 |
-
"split": "train",
|
| 119 |
-
"weight": 0.15
|
| 120 |
-
},
|
| 121 |
-
"conversational_data": {
|
| 122 |
-
"path": "sharegpt",
|
| 123 |
-
"split": "train",
|
| 124 |
-
"weight": 0.20
|
| 125 |
-
},
|
| 126 |
-
"multilingual_corpus": {
|
| 127 |
-
"path": "mc4",
|
| 128 |
-
"split": "train",
|
| 129 |
-
"weight": 0.15
|
| 130 |
-
}
|
| 131 |
-
},
|
| 132 |
"data_preprocessing": {
|
| 133 |
"max_seq_length": 131072,
|
| 134 |
"truncation": true,
|
|
|
|
| 1 |
{
|
| 2 |
+
"model_name_or_path": "DeepXR/Helion-2.5-Rnd",
|
| 3 |
"output_dir": "./checkpoints/helion-2.5-rnd",
|
| 4 |
"overwrite_output_dir": true,
|
| 5 |
"do_train": true,
|
|
|
|
| 102 |
"ddp_find_unused_parameters": false,
|
| 103 |
"ddp_bucket_cap_mb": null,
|
| 104 |
"ddp_broadcast_buffers": null,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
"data_preprocessing": {
|
| 106 |
"max_seq_length": 131072,
|
| 107 |
"truncation": true,
|