elliotthwang/KimLam-Ministral-4b-instruct-tw_train_outputs

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: ministral/Ministral-4b-instruct
+base_model: elliotthwang/Ministral-4b-instruct-tw_1k
 library_name: transformers
 model_name: outputs
 tags:
@@ -11,7 +11,7 @@ licence: license
 
 # Model Card for outputs
 
-This model is a fine-tuned version of [ministral/Ministral-4b-instruct](https://huggingface.co/ministral/Ministral-4b-instruct).
+This model is a fine-tuned version of [elliotthwang/Ministral-4b-instruct-tw_1k](https://huggingface.co/elliotthwang/Ministral-4b-instruct-tw_1k).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -35,7 +35,7 @@ This model was trained with SFT.
 ### Framework versions
 
 - TRL: 0.16.1
-- Transformers: 4.50.3
+- Transformers: 4.51.1
 - Pytorch: 2.6.0+cu124
 - Datasets: 3.5.0
 - Tokenizers: 0.21.1

adapter_config.json

@@ -1,7 +1,7 @@
 {
   "alpha_pattern": {},
   "auto_mapping": null,
-  "base_model_name_or_path": "ministral/Ministral-4b-instruct",
+  "base_model_name_or_path": "elliotthwang/Ministral-4b-instruct-tw_1k",
   "bias": "none",
   "eva_config": null,
   "exclude_modules": null,
@@ -25,9 +25,9 @@
   "target_modules": [
     "o_proj",
     "gate_proj",
-    "k_proj",
     "q_proj",
-    "v_proj"
+    "v_proj",
+    "k_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:324e3d36eb24c6e28cf873842f5de93a1e86af98452e05502246d1ba4846638f
+oid sha256:aacb936da69c016bd8269af3f4779c532230c31c50434b0eecf02d3db87e6d9a
 size 38839760

runs/Apr15_02-41-38_4d29867e47ff/events.out.tfevents.1744684899.4d29867e47ff.596.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db947fa4541953706f94806e92f7543fdb9e7e445711ebcd642380386320bbdf
+size 94471

tokenizer_config.json

@@ -37,7 +37,7 @@
   "legacy": false,
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "</s>",
-  "padding_side": "left",
+  "padding_side": "right",
   "sp_model_kwargs": {},
   "spaces_between_special_tokens": false,
   "tokenizer_class": "LlamaTokenizer",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7cf017de83701343051ddfd1b8bed347d97fa8b7c12ea0e9c6a0ab097a20cac3
-size 5624
+oid sha256:82404c1bf537fc27d98e14da82fbebf16f5c7bb3d532cadb36de4fc1fba53217
+size 5560