Model save

README.md

@@ -2,13 +2,9 @@
 license: other
 library_name: peft
 tags:
-- alignment-handbook
-- generated_from_trainer
 - trl
 - sft
 - generated_from_trainer
-datasets:
-- ruozhiba
 base_model: 01-ai/Yi-6B
 model-index:
 - name: Yi-6B-ruozhiba3
@@ -20,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Yi-6B-ruozhiba3
 
-This model is a fine-tuned version of [01-ai/Yi-6B](https://huggingface.co/01-ai/Yi-6B) on the ruozhiba dataset.
+This model is a fine-tuned version of [01-ai/Yi-6B](https://huggingface.co/01-ai/Yi-6B) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 3.9909
+- Loss: 4.3351
 
 ## Model description
 
@@ -48,7 +44,7 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 20
+- num_epochs: 3
 
 ### Training results
 
@@ -73,7 +69,6 @@ The following hyperparameters were used during training:
 | 0.0913        | 17.0  | 935  | 3.9921          |
 | 0.0895        | 18.0  | 990  | 3.9940          |
 | 0.0671        | 19.0  | 1045 | 3.9915          |
-| 0.0671        | 20.0  | 1100 | 3.9909          |
 
 
 ### Framework versions

adapter_config.json

@@ -19,12 +19,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "down_proj",
     "up_proj",
-    "k_proj",
     "o_proj",
-    "q_proj",
+    "down_proj",
     "gate_proj",
+    "q_proj",
+    "k_proj",
     "v_proj"
   ],
   "task_type": "CAUSAL_LM"

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7fd7f806620274b7515b5b677e684b071a7f8cdbd75fe8fb07253028745e9eb6
-size 72673912
+oid sha256:5db678411f47d7a4e0750ca5ac980c5058271f0e4bfcf34e49a24e258b5772fc
+size 145287696

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 20.0,
-    "eval_loss": 3.990852117538452,
-    "eval_runtime": 1.2395,
+    "eval_loss": 4.335062026977539,
+    "eval_runtime": 6.4371,
     "eval_samples": 23,
-    "eval_samples_per_second": 18.556,
-    "eval_steps_per_second": 4.841,
-    "train_loss": 0.6849299516461113,
-    "train_runtime": 917.6848,
+    "eval_samples_per_second": 3.573,
+    "eval_steps_per_second": 0.932,
+    "train_loss": 0.0,
+    "train_runtime": 10.2256,
     "train_samples": 217,
-    "train_samples_per_second": 4.729,
-    "train_steps_per_second": 1.199
+    "train_samples_per_second": 63.664,
+    "train_steps_per_second": 16.136
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "eval_loss": 3.990852117538452,
-    "eval_runtime": 1.2395,
+    "eval_loss": 4.335062026977539,
+    "eval_runtime": 6.4371,
     "eval_samples": 23,
-    "eval_samples_per_second": 18.556,
-    "eval_steps_per_second": 4.841
+    "eval_samples_per_second": 3.573,
+    "eval_steps_per_second": 0.932
 }

tokenizer_config.json

@@ -28,7 +28,7 @@
     }
   },
   "bos_token": "<|startoftext|>",
-  "chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}",
+  "chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|startoftext|><|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|startoftext|><|assistant|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|startoftext|><|assistant|>' }}\n{% endif %}\n{% endfor %}",
   "clean_up_tokenization_spaces": false,
   "eos_token": "<|endoftext|>",
   "legacy": true,

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "train_loss": 0.6849299516461113,
-    "train_runtime": 917.6848,
+    "train_loss": 0.0,
+    "train_runtime": 10.2256,
     "train_samples": 217,
-    "train_samples_per_second": 4.729,
-    "train_steps_per_second": 1.199
+    "train_samples_per_second": 63.664,
+    "train_steps_per_second": 16.136
 }

trainer_state.json

@@ -496,28 +496,20 @@
       "loss": 0.0671,
       "step": 1100
     },
-    {
-      "epoch": 20.0,
-      "eval_loss": 3.990852117538452,
-      "eval_runtime": 1.2506,
-      "eval_samples_per_second": 18.391,
-      "eval_steps_per_second": 4.798,
-      "step": 1100
-    },
     {
       "epoch": 20.0,
       "step": 1100,
       "total_flos": 3.807078373542298e+16,
-      "train_loss": 0.6849299516461113,
-      "train_runtime": 917.6848,
-      "train_samples_per_second": 4.729,
-      "train_steps_per_second": 1.199
+      "train_loss": 0.0,
+      "train_runtime": 10.2256,
+      "train_samples_per_second": 63.664,
+      "train_steps_per_second": 16.136
     }
   ],
   "logging_steps": 20,
-  "max_steps": 1100,
+  "max_steps": 165,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 20,
+  "num_train_epochs": 3,
   "save_steps": 20,
   "total_flos": 3.807078373542298e+16,
   "train_batch_size": 4,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3540153cddf246b726f4ba132c8c285fde88eb98729b0cff8eacff9abe1bffb4
+oid sha256:b591bed5d250490d702e6db47039b4083880a6eccfa19736ca1c3fe16762de21
 size 4728