Sync

README.md

@@ -221,6 +221,8 @@ Funding support
 
 NIST-MGI (https://www.nist.gov/mgi) and CHIPS (https://www.nist.gov/chips)
 
+Note: This project was originally developed under the github.com/usnistgov organization. New updates and developments will be carried out here.
+
 Code of conduct
 --------------------
 

atomgpt/examples/inverse_model/config.json

@@ -25,5 +25,6 @@
     "load_in_4bit": true,
     "instruction": "Below is a description of a superconductor material.",
     "alpaca_prompt": "### Instruction:\n{}\n### Input:\n{}\n### Output:\n{}",
-    "output_prompt": " Generate atomic structure description with lattice lengths, angles, coordinates and atom types."
+    "output_prompt": " Generate atomic structure description with lattice lengths, angles, coordinates and atom types.",
+    "hp_cfg_path": "hp_search_config.json"
 }

atomgpt/examples/inverse_model/hp_search_config.json

@@ -0,0 +1,25 @@
+{
+  "n_trials": 25,
+  "time_repeats": 1,
+  "objective_metric": "final_eval_loss",
+  "parameters": {
+    "learning_rate": { "type": "float", "low": 5e-6, "high": 5e-4, "log": true, "include": true },
+    "per_device_train_batch_size": { "type": "categorical", "choices": [1, 2, 4], "include": true },
+    "num_epochs": { "type": "int", "low": 1, "high": 5, "include": false },
+    "gradient_accumulation_steps": { "type": "categorical", "choices": [1, 2, 4, 8], "include": false },
+    "lora_rank": { "type": "categorical", "choices": [8, 16, 32, 64], "include": false },
+    "lora_alpha": { "type": "categorical", "choices": [8, 16, 32, 64], "include": false },
+    "max_seq_length": { "type": "categorical", "choices": [1024, 1536, 2048], "include": false },
+    "optim": { "type": "categorical", "choices": ["adamw_torch", "adamw_torch_fused", "adamw_bnb_8bit", "adamw_hf"], "include": false },
+    "lr_scheduler_type": { "type": "categorical", "choices": ["linear", "cosine", "cosine_with_restarts", "polynomial", "constant_with_warmup"], "include": false },
+    "warmup_ratio": { "type": "float", "low": 0.0, "high": 0.1, "include": false },
+    "logging_steps": { "type": "int", "low": 5, "high": 200, "include": false },
+    "seed_val": { "type": "int", "low": 1, "high": 10000, "include": false },
+    "dataset_num_proc": { "type": "categorical", "choices": [1, 2, 4, 8], "include": false },
+    "val_ratio": { "type": "float", "low": 0.05, "high": 0.2, "include": false },
+    "test_ratio": { "type": "float", "low": 0.1, "high": 0.3, "include": false },
+    "num_train": { "type": "int", "low": 0, "high": 1000000, "include": false },
+    "num_test": { "type": "int", "low": 0, "high": 100000, "include": false }
+  }
+}
+