Upload folder using huggingface_hub

ml_vir_mix1500_comp_only/adapter_config.json

@@ -0,0 +1,45 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "mistralai/Mistral-Large-Instruct-2411",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 128,
+  "lora_bias": false,
+  "lora_dropout": 0.1,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "v_proj",
+    "k_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

ml_vir_mix1500_comp_only/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c14d3d5091e0afc6fd2b32b89bc06768dc0caf9363760f654f7b5986918af786
+size 853638320

ml_vir_mix1500_comp_only/training_metadata.json

@@ -0,0 +1,37 @@
+{
+  "model_name": "ML-VIR-MIX1500-COMP-ONLY",
+  "model_id": "mistralai/Mistral-Large-Instruct-2411",
+  "dataset": "data/virtue_ethics_mix1500.jsonl",
+  "architecture": "1-stage mixed, COMPLETION-ONLY loss (prompt masked, no EOS)",
+  "loss_masking": "completion_only_prompt_masked_no_eos",
+  "train_samples": 1350,
+  "eval_samples": 150,
+  "lora_r": 64,
+  "lora_alpha": 128,
+  "lora_targets": [
+    "q_proj",
+    "v_proj",
+    "k_proj",
+    "o_proj"
+  ],
+  "lora_dropout": 0.1,
+  "epochs": 3,
+  "batch_size": 1,
+  "grad_accum": 8,
+  "lr": 0.0002,
+  "warmup_steps": 50,
+  "max_seq_length": 1024,
+  "quantization": "4bit_nf4",
+  "seed": 42,
+  "method": "self-teaching (Mistral Large generates + critiques + revises) \u2014 clean helpful base_cai slot",
+  "final_train_loss": 0.2718,
+  "final_eval_loss": 0.5732227563858032,
+  "best_eval_loss": 0.5732227563858032,
+  "lib_versions": {
+    "torch": "2.4.1+cu124",
+    "cuda": "12.4",
+    "transformers": "4.48.3",
+    "peft": "0.19.1",
+    "bitsandbytes": "0.49.0"
+  }
+}