Upload 6 files

README.md

@@ -0,0 +1,3 @@
+---
+library_name: peft
+---

adapter_config.json

@@ -0,0 +1,20 @@
+{
+  "base_model_name_or_path": "models\\elinas_llama-7b-hf-transformers-4.29",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 256,
+  "lora_dropout": 0.05,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 128,
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3cb63346badf925fa8b0b247743c962c1717bae09d773a4a9b5c636b1d981aca
+size 268481485

training_log.json

@@ -0,0 +1,16 @@
+{
+  "base_model_name": "elinas_llama-7b-hf-transformers-4.29",
+  "base_model_class": "LlamaForCausalLM",
+  "base_loaded_in_4bit": true,
+  "base_loaded_in_8bit": false,
+  "projections": "q, v",
+  "loss": 1.2101,
+  "learning_rate": 4.1443278256170227e-10,
+  "epoch": 1.99,
+  "current_steps": 15686,
+  "train_runtime": 47331.0727,
+  "train_samples_per_second": 0.67,
+  "train_steps_per_second": 0.005,
+  "total_flos": 6.648022708267254e+17,
+  "train_loss": 1.2976071742491992
+}

training_parameters.json

@@ -0,0 +1,30 @@
+{
+  "lora_name": "discordchat7b",
+  "always_override": true,
+  "save_steps": 1000.0,
+  "micro_batch_size": 2,
+  "batch_size": 128,
+  "epochs": 2.0,
+  "learning_rate": "1e-5",
+  "lr_scheduler_type": "cosine",
+  "lora_rank": 128,
+  "lora_alpha": 256,
+  "lora_dropout": 0.05,
+  "cutoff_len": 1024,
+  "dataset": "mini-10k",
+  "eval_dataset": "None",
+  "format": "vicuna-format",
+  "eval_steps": 500.0,
+  "raw_text_file": "formatted_messages",
+  "overlap_len": 64,
+  "newline_favor_len": 64,
+  "higher_rank_limit": false,
+  "warmup_steps": 100.0,
+  "optimizer": "adamw_bnb_8bit",
+  "hard_cut_string": "\\n\\n\\n",
+  "train_only_after": "",
+  "stop_at_loss": 0,
+  "add_eos_token": false,
+  "min_chars": 0.0,
+  "report_to": "wandb"
+}

training_prompt.json

@@ -0,0 +1,3 @@
+{
+  "template_type": "raw_text"
+}