Upload meta/step_0002000.json with huggingface_hub
Browse files- meta/step_0002000.json +66 -0
meta/step_0002000.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_val_loss": 3.6368123404681683,
|
| 3 |
+
"stream_offsets": {
|
| 4 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-1/general/train": 7419,
|
| 5 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-1/math/train": 7419,
|
| 6 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-1/code/train": 7419,
|
| 7 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-1/science/train": 7419,
|
| 8 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/math/train": 7419,
|
| 9 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/code/train": 7419,
|
| 10 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/science/train": 7419,
|
| 11 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/general/train": 7419,
|
| 12 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/tool_calling/train": 7419,
|
| 13 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/instruction-following/train": 7419,
|
| 14 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/swe_repair/train": 7419,
|
| 15 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/swe_localization/train": 7419,
|
| 16 |
+
"nvidia/Nemotron-Cascade-SFT-Stage-2/swe_testgen/train": 7419,
|
| 17 |
+
"nvidia/Nemotron-Cascade-2-SFT-Data/math/train": 7419,
|
| 18 |
+
"nvidia/Nemotron-Cascade-2-SFT-Data/science/train": 7419,
|
| 19 |
+
"nvidia/Nemotron-Cascade-2-SFT-Data/chat/train": 7419,
|
| 20 |
+
"nvidia/Nemotron-Cascade-2-SFT-Data/instruction_following/train": 7419,
|
| 21 |
+
"nvidia/Nemotron-Cascade-2-SFT-Data/safety/train": 3570,
|
| 22 |
+
"nvidia/Nemotron-Cascade-2-SFT-Data/conversational_agent/train": 7419,
|
| 23 |
+
"nvidia/Nemotron-Cascade-2-SFT-Data/swe/train": 7419,
|
| 24 |
+
"nvidia/Nemotron-Post-Training-Dataset-v1/default/chat": 7419,
|
| 25 |
+
"nvidia/Nemotron-Post-Training-Dataset-v1/default/code": 7419,
|
| 26 |
+
"nvidia/Nemotron-Post-Training-Dataset-v1/default/math": 7419,
|
| 27 |
+
"nvidia/Nemotron-Post-Training-Dataset-v1/default/stem": 7419,
|
| 28 |
+
"nvidia/Nemotron-Post-Training-Dataset-v1/default/tool_calling": 7419,
|
| 29 |
+
"nvidia/AceReason-1.1-SFT/default/train": 7419,
|
| 30 |
+
"nvidia/OpenMathInstruct-2/default/train": 7419,
|
| 31 |
+
"nvidia/OpenMathReasoning/default/cot": 7419,
|
| 32 |
+
"nvidia/OpenMathReasoning/default/tir": 7419,
|
| 33 |
+
"nvidia/OpenMathReasoning/default/genselect": 7419,
|
| 34 |
+
"nvidia/Nemotron-Math-v2/default/high_part00": 7419,
|
| 35 |
+
"nvidia/Nemotron-Math-v2/default/high_part01": 7419,
|
| 36 |
+
"nvidia/Nemotron-Math-v2/default/high_part02": 7419,
|
| 37 |
+
"nvidia/Nemotron-Math-v2/default/medium": 7419,
|
| 38 |
+
"nvidia/Nemotron-Math-v2/default/low": 7419,
|
| 39 |
+
"nvidia/Nemotron-SFT-Math-v3/default/train": 7418,
|
| 40 |
+
"nvidia/Nemotron-Research-GooseReason-0.7M/default/math": 0,
|
| 41 |
+
"nvidia/Nemotron-Research-GooseReason-0.7M/default/code": 0,
|
| 42 |
+
"nvidia/Nemotron-Research-GooseReason-0.7M/default/stem": 0,
|
| 43 |
+
"nvidia/OpenCodeGeneticInstruct/mixtral-8x22b-instruct/train": 7418,
|
| 44 |
+
"nvidia/OpenCodeGeneticInstruct/qwen2.5-32b-instruct/train": 7418,
|
| 45 |
+
"nvidia/OpenCodeReasoning/split_0/split_0": 7418,
|
| 46 |
+
"nvidia/OpenCodeReasoning/split_1/split_1": 7418,
|
| 47 |
+
"nvidia/Nemotron-Math-HumanReasoning/default/train": 250,
|
| 48 |
+
"nvidia/Nemotron-SFT-Agentic-v2/default/interactive_agent": 7418,
|
| 49 |
+
"nvidia/Nemotron-SFT-Agentic-v2/default/tool_calling": 0,
|
| 50 |
+
"nvidia/Nemotron-Instruction-Following-Chat-v1/default/chat_if": 7418,
|
| 51 |
+
"nvidia/Nemotron-Instruction-Following-Chat-v1/default/structured_outputs": 4969,
|
| 52 |
+
"nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_python_part00": 7418,
|
| 53 |
+
"nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_python_part01": 7418,
|
| 54 |
+
"nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_cpp_part00": 7418,
|
| 55 |
+
"nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_cpp_part01": 7418,
|
| 56 |
+
"nvidia/Nemotron-RL-Super-Training-Blends/default/rlvr1": 0
|
| 57 |
+
},
|
| 58 |
+
"avg_train_loss": 2.226964343070984,
|
| 59 |
+
"sft_step": 2000,
|
| 60 |
+
"timestamp": "2026-04-20 08:49:37",
|
| 61 |
+
"global_batch": 128,
|
| 62 |
+
"seq_len": 8192,
|
| 63 |
+
"lr_start": 5e-05,
|
| 64 |
+
"lr_min": 1e-06,
|
| 65 |
+
"vocab_size": 32777
|
| 66 |
+
}
|