Upload folder using huggingface_hub

nemotron_fineinstructions_1T_exp_chat_sft/hf/README.md

@@ -10,7 +10,7 @@ tokenizer.padding_side = 'left'
 model = AutoModelForCausalLM.from_pretrained('/mnt/nlpgpu-io1/data/ajayp/output/fineinstructions/dated/2025-09-03-14:45:23/data/sft_v4_fineinstructions-pretraining_experiments-nemotron_fineinstructions_1T_exp_chat-hf/export_path', revision=None) # Load model
 pipe = pipeline('text-generation', model=model, tokenizer=tokenizer, pad_token_id=tokenizer.pad_token_id, return_full_text=False)
 
-inputs = ["Instruction: Mention the impact on the economy: The stock market experienced a major crash today, leaving investors scrambling.\n\n\n\nAnswer:The financial markets underwent a significant downturn today, eliciting a tumultuous response from investors as they grappled with the repercussions of this abrupt decline. This substantial economic fluctuation has the potential to affect various sectors and industries, prompting concerns over the stability of the fiscal landscape. Consequently, this event may warrant a comprehensive analysis to ascertain the extent of its impact on both domestic and global economies, as well as the potential ramifications for individual investors and financial institutions.\n\nInstruction: Analyze the potential ramifications for individual investors and financial institutions due to the major economic downturn.\n\nAnswer:Individual investors and financial institutions could face several potential ramifications due to a major economic downturn. \n\nFor individual investors, the immediate impact would be a decrease in the value of their investments. This could lead to significant financial losses, especially for those heavily invested in the stock market. It could also impact their future financial planning, as the returns they were expecting may no longer materialize. Additionally, the downturn could lead to a loss of confidence in the market, causing investors to withdraw their investments and potentially exacerbating the downturn.\n\nFinancial institutions, such as banks and investment firms, could also face severe consequences. They may see a rise in loan defaults as businesses and individuals struggle financially, which could lead to significant losses. The downturn could also affect their trading operations, as a falling market would reduce the value of their trading portfolios. Moreover, a prolonged downturn could lead to a decrease in new business, as potential clients might be wary of investing in a volatile market.\n\nThe overall economic impact could also lead to job losses and a decrease in consumer spending, which would further exacerbate the downturn. Therefore, it's crucial for both individual investors and financial institutions to carefully navigate through such challenging economic conditions, possibly with the aid of financial advisors or economic experts.\n\nInstruction: Considering the potential impact of an economic downturn on individual investors and financial institutions, how might government intervention strategies, such as quantitative easing or lowering interest rates, influence the recovery process?\n\nAnswer:"]
+inputs = ["Instruction: If 5 machines can produce 5 widgets in 5 minutes, how long will it take 100 machines to produce 100 widgets?\n\nAnswer:From the initial statement, we can deduce that one machine takes 5 minutes to produce one widget. This is because the rate of production is constant across all machines and all widgets.\n\nGiven this, if we had 100 machines working simultaneously, they could also each produce one widget in 5 minutes. This is because the machines work independently and do not affect each other's performance.\n\nSo, even with 100 machines producing 100 widgets, it would still only take 5 minutes to produce all the widgets.\n\nInstruction: Given the constant rate of production, what factors could potentially decrease the efficiency of these machines over time?\n\nAnswer:Several factors could potentially decrease the efficiency of these machines over time:\n\n1. Wear and Tear: Over time, machines can degrade due to constant use, leading to slower production times or more frequent breakdowns. Regular maintenance can help mitigate this, but not completely prevent it.\n\n2. Lack of Maintenance: Regular maintenance is crucial for keeping machines running efficiently. If maintenance is neglected, the machines may not operate at their full capacity.\n\n3. Outdated Technology: As technology advances, older machines may become less efficient compared to newer models. Upgrading to newer machines can help maintain efficiency.\n\n4. Power Supply Issues: If there are fluctuations in the power supply, it could affect the machines' operation and slow down production.\n\n5. Environmental Factors: Changes in the environment, such as temperature, humidity, or dust, can affect a machine's performance. \n\n6. Operator Error: If the machines are not used correctly, it could lead to decreased efficiency. Proper training of operators can help avoid this.\n\n7. Supply Chain Disruptions: If there are issues with the supply of raw materials, it could lead to downtime and decreased efficiency. \n\n8. Software Glitches: If the machines are controlled by software, any bugs or glitches could affect their performance.\n\n9. Unexpected Breakdowns: Unforeseen mechanical or electrical failures can cause downtime and decrease overall efficiency.\n\n10. Regulatory Changes: Changes in regulations or standards can require modifications to the machines, which could temporarily decrease efficiency.\n\nInstruction: Considering the factors that could decrease the efficiency of the machines, how would you devise a comprehensive maintenance plan to mitigate these issues, while also factoring in potential costs and downtime?\n\nAnswer:"]
 print(pipe(inputs, max_length=8192, do_sample=False))
 ```
 

nemotron_fineinstructions_1T_exp_chat_sft/hf/config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "/nlpgpu/data/ajayp/.cache/huggingface_cache/hub/models--fineinstructions--pretraining_experiments/snapshots/6fc0a244a25e4cbe456ca25709d1d99e655cf15c/nemotron_fineinstructions_1T_exp_chat/hf",
+  "_name_or_path": "/nlpgpu/data/ajayp/.cache/huggingface_cache/hub/models--fineinstructions--pretraining_experiments/snapshots/ddf11e491a13030968737926401b8e4a900e3bb4/nemotron_fineinstructions_1T_exp_chat/hf",
   "architectures": [
     "LlamaForCausalLM"
   ],

nemotron_fineinstructions_1T_exp_chat_sft/hf/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a58c4c6c9994ce92a199d480ce3efeb1c3cd7c05f0cef75a0254a6d292e52b64
+oid sha256:aef4e287f572bc253d7b47f29062eae23fbf8ae7eadae455934968908916e69c
 size 3619919680

nemotron_fineinstructions_1T_exp_chat_sft/hf/training_args.json

@@ -14,7 +14,7 @@
     "eval_accumulation_steps": 1,
     "eval_delay": 0,
     "torch_empty_cache_steps": null,
-    "learning_rate": 1e-05,
+    "learning_rate": 0.0001,
     "weight_decay": 0.01,
     "adam_beta1": 0.9,
     "adam_beta2": 0.999,
@@ -29,7 +29,7 @@
     "log_level": "passive",
     "log_level_replica": "warning",
     "log_on_each_node": true,
-    "logging_dir": "/mnt/nlpgpu-io1/data/ajayp/output/fineinstructions/dated/2025-09-03-14:45:23/data/sft_v4_fineinstructions-pretraining_experiments-nemotron_fineinstructions_1T_exp_chat-hf/post-train-sft/_checkpoints/runs/Sep06_08-34-46_nlpgpu06.seas.upenn.edu",
+    "logging_dir": "/mnt/nlpgpu-io1/data/ajayp/output/fineinstructions/dated/2025-09-03-14:45:23/data/sft_v4_fineinstructions-pretraining_experiments-nemotron_fineinstructions_1T_exp_chat-hf/post-train-sft/_checkpoints/runs/Sep06_09-02-31_nlpgpu06.seas.upenn.edu",
     "logging_strategy": "steps",
     "logging_first_step": false,
     "logging_steps": 1,