Model save

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
+base_model: Qwen/Qwen2-0.5B-Instruct
 library_name: peft
 license: apache-2.0
 tags:
@@ -16,7 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # output
 
-This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-Chat-v1.0](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0) on an unknown dataset.
+This model is a fine-tuned version of [Qwen/Qwen2-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2-0.5B-Instruct) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.0816
 
 ## Model description
 
@@ -35,20 +37,27 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 2e-05
-- train_batch_size: 2
+- learning_rate: 0.002
+- train_batch_size: 1
 - eval_batch_size: 8
 - seed: 42
-- gradient_accumulation_steps: 4
+- gradient_accumulation_steps: 8
 - total_train_batch_size: 8
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - num_epochs: 1
 
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 1.2506        | 1.0   | 1250 | 1.0816          |
+
+
 ### Framework versions
 
 - PEFT 0.12.0
-- Transformers 4.44.0.dev0
+- Transformers 4.42.4
 - Pytorch 2.3.1+cu121
 - Datasets 2.20.0
 - Tokenizers 0.19.1

adapter_config.json

@@ -1,7 +1,7 @@
 {
   "alpha_pattern": {},
   "auto_mapping": null,
-  "base_model_name_or_path": "Qwen/Qwen2-7B-Instruct",
+  "base_model_name_or_path": "Qwen/Qwen2-0.5B-Instruct",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,
@@ -11,19 +11,19 @@
   "layers_to_transform": null,
   "loftq_config": {},
   "lora_alpha": 16,
-  "lora_dropout": 0.1,
+  "lora_dropout": 0.3,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 64,
+  "r": 16,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
     "k_proj",
     "q_proj",
-    "o_proj"
+    "o_proj",
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1285584ced1cd1b12189936f47cb41f863ae49cb52d05232dda4ae5426ed7de7
-size 161510984
+oid sha256:df15b611e7e64a99c4de4ce6013fc7c74556a6da6463b4fa3dac25756b3e0c84
+size 8676008

tokenizer.json

@@ -2,7 +2,7 @@
   "version": "1.0",
   "truncation": {
     "direction": "Right",
-    "max_length": 4096,
+    "max_length": 256,
     "strategy": "LongestFirst",
     "stride": 0
   },

tokenizer_config.json

@@ -35,7 +35,7 @@
   "clean_up_tokenization_spaces": false,
   "eos_token": "<|im_end|>",
   "errors": "replace",
-  "model_max_length": 131072,
+  "model_max_length": 32768,
   "pad_token": "<|im_end|>",
   "split_special_tokens": false,
   "tokenizer_class": "Qwen2Tokenizer",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b6091982a7044ee74783c3b4e9a883f4c5599f28978dbf6099fb1e7213119dfd
-size 5368
+oid sha256:18dcabf791b4cb3f0ad098a7aa11d08ffae4dce95bc8aef016dff302eb884c49
+size 5304