TymofiiNasobko/Lapa-function-calling

README.md

@@ -4,8 +4,8 @@ library_name: transformers
 model_name: Lapa-function-calling
 tags:
 - generated_from_trainer
-- trl
 - sft
+- trl
 licence: license
 ---
 
@@ -36,7 +36,7 @@ This model was trained with SFT.
 
 - TRL: 0.25.1
 - Transformers: 4.57.3
-- Pytorch: 2.8.0+cu129
+- Pytorch: 2.6.0+cu118
 - Datasets: 4.4.1
 - Tokenizers: 0.22.1
 

adapter_config.json

@@ -21,7 +21,10 @@
   "lora_dropout": 0.05,
   "megatron_config": null,
   "megatron_core": "megatron.core",
-  "modules_to_save": null,
+  "modules_to_save": [
+    "lm_head",
+    "embed_tokens"
+  ],
   "peft_type": "LORA",
   "peft_version": "0.18.0",
   "qalora_group_size": 16,
@@ -29,16 +32,16 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "fc2",
     "o_proj",
-    "k_proj",
-    "gate_proj",
+    "v_proj",
+    "down_proj",
+    "fc2",
     "q_proj",
+    "out_proj",
     "fc1",
     "up_proj",
-    "down_proj",
-    "out_proj",
-    "v_proj"
+    "gate_proj",
+    "k_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b51c582896766642a149c855746afbcb3818bd27cda03655a6cd4de69ae656e8
-size 4746766832
+oid sha256:d5bd45b8610d7433d4628cc89c58265703031a5b658a642c73bf3bb7de3c0ffd
+size 8774281960

chat_template.jinja

@@ -42,14 +42,14 @@
     {%- else -%}
         {{ raise_exception("Invalid content type") }}
     {%- endif -%}
-    {%- if role == "model" -%}
+{%- if role == "model" -%}
 {% generation %}
 {{ '<end_of_turn>
 ' }}{% endgeneration %}
-    {%- else -%}
+{%- else -%}
 {{ '<end_of_turn>
 ' }}
-    {%- endif -%}
+{%- endif -%}
 {%- endfor -%}
 
 {%- if add_generation_prompt -%}

special_tokens_map.json

@@ -8,7 +8,7 @@
     "single_word": false
   },
   "eoi_token": "<end_of_image>",
-  "eos_token": "<eos>",
+  "eos_token": "<end_of_turn>",
   "image_token": "<image_soft_token>",
   "pad_token": {
     "content": "<pad>",

tokenizer_config.json

@@ -2207,7 +2207,7 @@
   "bos_token": "<bos>",
   "clean_up_tokenization_spaces": false,
   "eoi_token": "<end_of_image>",
-  "eos_token": "<eos>",
+  "eos_token": "<end_of_turn>",
   "extra_special_tokens": {
     "boi_token": "<start_of_image>",
     "eoi_token": "<end_of_image>",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:264c88be027ded39b70324149dbe1cf4b50e33e90c1a743861991a753109baff
-size 6225
+oid sha256:30e1c54445c5beccd239bb6cc38804b3a0439fb2dae5c02ee756d2bb97e018e9
+size 5816