Continued pretraining: FineWeb-Edu + rewrite6 byte hybrid + no-prompt-15k
Browse files
continued_pretraining_manifest.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model": "appvoid/void.0",
|
| 3 |
+
"sources": {
|
| 4 |
+
"cache_version": 3,
|
| 5 |
+
"rewrite6_source": "appvoid/rewrite6",
|
| 6 |
+
"rewrite6_split": "train",
|
| 7 |
+
"fineweb_dataset": "HuggingFaceFW/fineweb-edu",
|
| 8 |
+
"fineweb_config": "sample-10BT",
|
| 9 |
+
"fineweb_split": "train",
|
| 10 |
+
"no_prompt_dataset": "appvoid/no-prompt-15k",
|
| 11 |
+
"no_prompt_split": "train",
|
| 12 |
+
"fields": {
|
| 13 |
+
"rewrite6": [
|
| 14 |
+
"instruction",
|
| 15 |
+
"text",
|
| 16 |
+
"output"
|
| 17 |
+
],
|
| 18 |
+
"fineweb_edu": [
|
| 19 |
+
"text"
|
| 20 |
+
],
|
| 21 |
+
"no_prompt_15k": [
|
| 22 |
+
"text"
|
| 23 |
+
]
|
| 24 |
+
},
|
| 25 |
+
"counts": {
|
| 26 |
+
"rewrite6": 205404,
|
| 27 |
+
"fineweb_edu": 205404,
|
| 28 |
+
"no_prompt_15k": 15000
|
| 29 |
+
},
|
| 30 |
+
"seed": 3407
|
| 31 |
+
},
|
| 32 |
+
"packing": {
|
| 33 |
+
"signature": "2a7abc0abb945816678da7dce3e6df927fc40e60c55d56ea18972abfd374e20a",
|
| 34 |
+
"block_size": 4096,
|
| 35 |
+
"blocks": 88660,
|
| 36 |
+
"real_tokens": 363149667,
|
| 37 |
+
"examples_seen": {
|
| 38 |
+
"rewrite6": 205404,
|
| 39 |
+
"fineweb_edu": 205404,
|
| 40 |
+
"no_prompt_15k": 15000
|
| 41 |
+
},
|
| 42 |
+
"source_token_counts": {
|
| 43 |
+
"rewrite6": 116587156,
|
| 44 |
+
"fineweb_edu": 241348518,
|
| 45 |
+
"no_prompt_15k": 5213993
|
| 46 |
+
},
|
| 47 |
+
"source_counts": {
|
| 48 |
+
"rewrite6": 205404,
|
| 49 |
+
"fineweb_edu": 205404,
|
| 50 |
+
"no_prompt_15k": 15000
|
| 51 |
+
},
|
| 52 |
+
"token_dtype": "uint32",
|
| 53 |
+
"length_dtype": "uint16",
|
| 54 |
+
"two_epoch_tokens": 726299334.0
|
| 55 |
+
},
|
| 56 |
+
"epochs": 2.0,
|
| 57 |
+
"max_context_length": 4096,
|
| 58 |
+
"learning_rate": 5e-05,
|
| 59 |
+
"micro_batch_size": 32,
|
| 60 |
+
"gradient_accumulation_steps": 1,
|
| 61 |
+
"precision": "bf16",
|
| 62 |
+
"attention_implementation": "sdpa",
|
| 63 |
+
"rewrite_layout": "instruction\\n\\n\"text bytes\"\\n\\n\"output bytes\""
|
| 64 |
+
}
|