Upload training_meta.json with huggingface_hub

training_meta.json

@@ -1,64 +1,64 @@
 {
   "best_val_loss": 1.6575258653610945,
   "stream_offsets": {
-    "nvidia/Nemotron-Cascade-SFT-Stage-1/general/train": 27416,
-    "nvidia/Nemotron-Cascade-SFT-Stage-1/math/train": 27416,
-    "nvidia/Nemotron-Cascade-SFT-Stage-1/code/train": 27416,
-    "nvidia/Nemotron-Cascade-SFT-Stage-1/science/train": 27416,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/math/train": 27416,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/code/train": 27415,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/science/train": 27415,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/general/train": 27415,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/tool_calling/train": 27415,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/instruction-following/train": 27415,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/swe_repair/train": 27415,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/swe_localization/train": 27415,
-    "nvidia/Nemotron-Cascade-SFT-Stage-2/swe_testgen/train": 27415,
-    "nvidia/Nemotron-Cascade-2-SFT-Data/math/train": 27415,
-    "nvidia/Nemotron-Cascade-2-SFT-Data/science/train": 27415,
-    "nvidia/Nemotron-Cascade-2-SFT-Data/chat/train": 27415,
-    "nvidia/Nemotron-Cascade-2-SFT-Data/instruction_following/train": 27415,
+    "nvidia/Nemotron-Cascade-SFT-Stage-1/general/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-1/math/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-1/code/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-1/science/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/math/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/code/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/science/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/general/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/tool_calling/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/instruction-following/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/swe_repair/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/swe_localization/train": 29711,
+    "nvidia/Nemotron-Cascade-SFT-Stage-2/swe_testgen/train": 29710,
+    "nvidia/Nemotron-Cascade-2-SFT-Data/math/train": 29710,
+    "nvidia/Nemotron-Cascade-2-SFT-Data/science/train": 29710,
+    "nvidia/Nemotron-Cascade-2-SFT-Data/chat/train": 29710,
+    "nvidia/Nemotron-Cascade-2-SFT-Data/instruction_following/train": 29710,
     "nvidia/Nemotron-Cascade-2-SFT-Data/safety/train": 3570,
-    "nvidia/Nemotron-Cascade-2-SFT-Data/conversational_agent/train": 27415,
-    "nvidia/Nemotron-Cascade-2-SFT-Data/swe/train": 27415,
-    "nvidia/Nemotron-Post-Training-Dataset-v1/default/chat": 27415,
-    "nvidia/Nemotron-Post-Training-Dataset-v1/default/code": 27415,
-    "nvidia/Nemotron-Post-Training-Dataset-v1/default/math": 27415,
-    "nvidia/Nemotron-Post-Training-Dataset-v1/default/stem": 27415,
-    "nvidia/Nemotron-Post-Training-Dataset-v1/default/tool_calling": 27415,
-    "nvidia/AceReason-1.1-SFT/default/train": 27415,
+    "nvidia/Nemotron-Cascade-2-SFT-Data/conversational_agent/train": 29710,
+    "nvidia/Nemotron-Cascade-2-SFT-Data/swe/train": 29710,
+    "nvidia/Nemotron-Post-Training-Dataset-v1/default/chat": 29710,
+    "nvidia/Nemotron-Post-Training-Dataset-v1/default/code": 29710,
+    "nvidia/Nemotron-Post-Training-Dataset-v1/default/math": 29710,
+    "nvidia/Nemotron-Post-Training-Dataset-v1/default/stem": 29710,
+    "nvidia/Nemotron-Post-Training-Dataset-v1/default/tool_calling": 29710,
+    "nvidia/AceReason-1.1-SFT/default/train": 29710,
     "nvidia/OpenMathInstruct-2/default/train": 58705,
     "nvidia/OpenMathReasoning/default/cot": 0,
     "nvidia/OpenMathReasoning/default/tir": 0,
     "nvidia/OpenMathReasoning/default/genselect": 0,
-    "nvidia/Nemotron-Math-v2/default/high_part00": 27415,
-    "nvidia/Nemotron-Math-v2/default/high_part01": 27415,
-    "nvidia/Nemotron-Math-v2/default/high_part02": 27415,
-    "nvidia/Nemotron-Math-v2/default/medium": 27415,
-    "nvidia/Nemotron-Math-v2/default/low": 27415,
-    "nvidia/Nemotron-SFT-Math-v3/default/train": 27415,
-    "nvidia/Nemotron-Research-GooseReason-0.7M/default/math": 27415,
-    "nvidia/Nemotron-Research-GooseReason-0.7M/default/code": 27415,
-    "nvidia/Nemotron-Research-GooseReason-0.7M/default/stem": 27415,
-    "nvidia/OpenCodeGeneticInstruct/mixtral-8x22b-instruct/train": 27415,
-    "nvidia/OpenCodeGeneticInstruct/qwen2.5-32b-instruct/train": 27415,
-    "nvidia/OpenCodeReasoning/split_0/split_0": 27415,
-    "nvidia/OpenCodeReasoning/split_1/split_1": 27415,
+    "nvidia/Nemotron-Math-v2/default/high_part00": 29710,
+    "nvidia/Nemotron-Math-v2/default/high_part01": 29710,
+    "nvidia/Nemotron-Math-v2/default/high_part02": 29710,
+    "nvidia/Nemotron-Math-v2/default/medium": 29710,
+    "nvidia/Nemotron-Math-v2/default/low": 29710,
+    "nvidia/Nemotron-SFT-Math-v3/default/train": 29710,
+    "nvidia/Nemotron-Research-GooseReason-0.7M/default/math": 29710,
+    "nvidia/Nemotron-Research-GooseReason-0.7M/default/code": 29710,
+    "nvidia/Nemotron-Research-GooseReason-0.7M/default/stem": 29710,
+    "nvidia/OpenCodeGeneticInstruct/mixtral-8x22b-instruct/train": 29710,
+    "nvidia/OpenCodeGeneticInstruct/qwen2.5-32b-instruct/train": 29710,
+    "nvidia/OpenCodeReasoning/split_0/split_0": 29710,
+    "nvidia/OpenCodeReasoning/split_1/split_1": 29710,
     "nvidia/Nemotron-Math-HumanReasoning/default/train": 0,
-    "nvidia/Nemotron-SFT-Agentic-v2/default/interactive_agent": 27415,
+    "nvidia/Nemotron-SFT-Agentic-v2/default/interactive_agent": 29710,
     "nvidia/Nemotron-SFT-Agentic-v2/default/tool_calling": 0,
-    "nvidia/Nemotron-Instruction-Following-Chat-v1/default/chat_if": 27415,
+    "nvidia/Nemotron-Instruction-Following-Chat-v1/default/chat_if": 29710,
     "nvidia/Nemotron-Instruction-Following-Chat-v1/default/structured_outputs": 4969,
-    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_python_part00": 27415,
-    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_python_part01": 27415,
-    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_cpp_part00": 27415,
-    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_cpp_part01": 27415,
+    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_python_part00": 29710,
+    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_python_part01": 29710,
+    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_cpp_part00": 29710,
+    "nvidia/Nemotron-Competitive-Programming-v1/default/competitive_coding_cpp_part01": 29710,
     "nvidia/Nemotron-RL-Super-Training-Blends/default/rlvr1": 3933,
-    "ibivibiv/math_instruct/default/train": 7005506,
-    "javi22/simple-instruct-dataset/default/train": 7005505,
+    "ibivibiv/math_instruct/default/train": 7007801,
+    "javi22/simple-instruct-dataset/default/train": 7007800,
     "agentlans/chatgpt/default/train": 0,
     "fineinstructions-pretraining/nemotron_fineinstructions_1T_exp_chat/default/train": 0,
-    "agentlans/chatgpt/default/train::hf://datasets/agentlans/chatgpt/HydraLM-airoboros-gpt4-1.4_alpaca.jsonl.zst": 27415,
+    "agentlans/chatgpt/default/train::hf://datasets/agentlans/chatgpt/HydraLM-airoboros-gpt4-1.4_alpaca.jsonl.zst": 29710,
     "agentlans/chatgpt/default/train::hf://datasets/agentlans/chatgpt/Mxode-Magpie-Pro-10K-GPT4o-mini.jsonl.zst": 0,
     "agentlans/chatgpt/default/train::hf://datasets/agentlans/chatgpt/PKU-Alignment-Align-Anything-Instruction-100K.jsonl.zst": 0,
     "agentlans/chatgpt/default/train::hf://datasets/agentlans/chatgpt/PawanKrd-gpt-4o-200k.jsonl.zst": 0,
@@ -116,9 +116,9 @@
     "agentlans/chatgpt/default/train::hf://datasets/agentlans/chatgpt/microsoft-orca-agentinstruct-1m-v1.jsonl.zst": 0,
     "agentlans/chatgpt/default/train::hf://datasets/agentlans/chatgpt/microsoft-orca-math-word-problems-200k.jsonl.zst": 0
   },
-  "avg_train_loss": 2.226559741258621,
-  "sft_step": 28500,
-  "timestamp": "2026-04-22 14:09:16",
+  "avg_train_loss": 2.22396332859993,
+  "sft_step": 29000,
+  "timestamp": "2026-04-22 14:42:21",
   "global_batch": 128,
   "seq_len": 8192,
   "lr_start": 5e-05,