初始化项目，由ModelHub XC社区提供模型

Model: nwdxlgzs/XL-LuaCopilot-1.7B-FFT-GGUF
Source: Original Platform

.gitattributes

@@ -0,0 +1,59 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-BF16.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-F16.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-IQ3_S.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q3_K.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q4_1.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q4_K.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q5_0.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q5_1.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q5_K.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text
+XL-LuaCopilot-1.7B-FFT-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,141 @@
+---
+tags:
+- unsloth
+- lua
+base_model:
+- nwdxlgzs/XL-LuaCopilot-1.7B-FFT-checkpoint-46000
+license: gpl-3.0
+library_name: transformers
+pipeline_tag: text-generation
+---
+# XL-LuaCopilot-1.7B-FFT
+
+XL-LuaCopilot-1.7B-FFT is a large language model (LLM) based on the Qwen architecture(Qwen3-1.7B), specifically designed for code generation tasks in Lua programming language. It has been full fine-tuned (FFT) to improve its performance and efficiency when generating Lua code.
+
+I sugggest you use `"chat_template_kwargs": {"enable_thinking": false}` because my train data with none thinking. I also found low `temperature` ususually works well for code generation tasks.
+
+quantize=["Q4_0", "Q4_1", "Q5_0", "Q5_1", "IQ3_XXS", "IQ3_S", "IQ3_M", "Q3_K", "IQ3_XS", "Q3_K_S", "Q3_K_M", "Q3_K_L", "IQ4_NL", "IQ4_XS", "Q4_K", "Q4_K_S", "Q4_K_M", "Q5_K", "Q5_K_S", "Q5_K_M", "Q6_K", "Q8_0", "F16", "BF16"]
+
+> `checkpoint-37000` is the checkpoint where the model had just entered the plateau phase with a lower loss, and it might be better than `checkpoint-46000`. However, the GGUF files I provide will still be released based on the final checkpoint at the end of training.
+
+## Train Samples
+
+1472000 (steps=46000)x(per_device_train_batch_size=8)x(gradient_accumulation_steps=4)x(device=1)
+
+datasets: 1464339 (luafiles=488113)x(split=3)
+
+epoch‌s = 1.005
+
+## How To Use
+
+> With OpenAI Compatible  API (llama.cpp:llama-server)
+
+```json
+-> REQUEST ->
+
+{
+    "model": "XL-LuaCopilot-1.7B-FFT",
+    "messages": [
+        {"role": "system","content": "prefix"},
+        {"role": "user","content": "do\n--打印：你好世界\n  local tex"},
+        {"role": "system","content": "suffix"},
+        {"role": "user","content": "nd"},
+        {"role": "system","content": "middle"}
+    ],
+    "stream": false,
+    "cache_prompt": false,
+    "samplers": "edkypmxt",
+    "temperature": 0.2,
+    "dynatemp_range": 0.1,
+    "dynatemp_exponent": 1,
+    "top_k": 20,
+    "top_p": 0.9,
+    "min_p": 0.05,
+    "typical_p": 1,
+    "xtc_probability": 0,
+    "xtc_threshold": 0.1,
+    "repeat_last_n": 32,
+    "repeat_penalty": 1.1,
+    "presence_penalty": 0,
+    "frequency_penalty": 0.5,
+    "dry_multiplier": 0,
+    "dry_base": 1.75,
+    "dry_allowed_length": 2,
+    "dry_penalty_last_n": -1,
+    "max_tokens": -1,
+    "timings_per_token": true,
+    "chat_template_kwargs": {"enable_thinking": false}
+}
+
+-> RESPONSE ->
+
+{
+    "choices": [
+        {
+            "finish_reason": "stop",
+            "index": 0,
+            "message": {
+                "role": "assistant",
+                "content": "<think>\n\n</think>\n\nt = \"你好世界\"\n  print(text)\ne"
+            }
+        }
+    ],
+    ...
+}
+```
+
+> I know Qwen has `<|fim_prefix|>` / `<|fim_suffix|>` / `<|fim_middle|>` tokens, but I'm not sure Qwen3 trains these tokens (I just know Qwen2.5-Coder does). To use code generation easily, I use chatml format.
+
+> If you just want to chat with it, you can use some tricks like this:
+```
+<|im_end|>
+<|im_start|>system
+prefix<|im_end|>
+<|im_start|>user
+do
+    --打印：你好世界
+    local tex<|im_end|>
+<|im_start|>system
+suffix<|im_end|>
+<|im_start|>user
+nd<|im_end|>
+<|im_start|>system
+middle
+```
+
+It dosen't work very well, but it's a good way let you fast try. It will convert to this prompt text:
+
+```
+<|im_start|>user
+<|im_end|>
+<|im_start|>system
+prefix<|im_end|>
+<|im_start|>user
+do
+    --打印：你好世界
+    local tex<|im_end|>
+<|im_start|>system
+suffix<|im_end|>
+<|im_start|>user
+nd<|im_end|>
+<|im_start|>system
+middle<|im_end|>
+```
+Hope model skip first `<|im_start|>user\n<|im_end|>` part.
+
+# Train Device
+
+> Online GPU is Expensive !
+
+| 类别           | 配置详情                                           |
+|----------------|---------------------------------------------------|
+| **镜像**       | Ubuntu 22.04                                      |
+| **PyTorch**    | 2.5.1                                             |
+| **Python**     | 3.12                                              |
+| **CUDA**       | 12.4                                              |
+| **GPU**        | RTX 4090 (24GB) * 1                               |
+| **CPU**        | 25 vCPU Intel(R) Xeon(R) Platinum 8481C           |
+| **内存**       | 90GB                                              |
+| **硬盘**       | 30 GB + 50 GB                                     |
+| **时长**       | 3 Day                                             |
+

XL-LuaCopilot-1.7B-FFT-BF16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:680d40cbecc8177f2901be94c845bd0bf578b670d9f5250ef38f4d72e2c2ada1
+size 3447344704

XL-LuaCopilot-1.7B-FFT-F16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d07e8a926af4d363f515ccb55b875852d151ccd00b5b04452f7c6bad5e74a7eb
+size 3447344704

XL-LuaCopilot-1.7B-FFT-IQ3_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d33d7a07947b2e6ad28d1e8cfddeb7dc6131c700e214f85e12549816a149185
+size 895657536

XL-LuaCopilot-1.7B-FFT-IQ3_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bda2e01db0a5db5c22150677a16de6b477c40fb4ad3b063245ac493711520b6f
+size 867247680

XL-LuaCopilot-1.7B-FFT-IQ3_XS.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e5532034d76d5607be3ea39d841c3e063487409f2f3d91aa70e6c9647dee058
+size 834217536

XL-LuaCopilot-1.7B-FFT-IQ3_XXS.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c7687d47e32f149f8656a0e247a1dd11f83eadbd12337cb391eda3aa132531b
+size 219881024

XL-LuaCopilot-1.7B-FFT-IQ4_NL.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dacc5b6868ccb2cc7a856acb5e3a1b100736c1736777e6c27f75a45713c4eba
+size 1059137088

XL-LuaCopilot-1.7B-FFT-IQ4_XS.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48024c6a36ceb2ef2289a78d7a144f781acb459bde30842b1614655f8bc8d530
+size 1016276544

XL-LuaCopilot-1.7B-FFT-Q3_K.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d06ab461534367d70e6161938ae3e173e05db61ca3ec19e048743def920aa1a
+size 939533888

XL-LuaCopilot-1.7B-FFT-Q3_K_L.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:671f230537a1c9b9020ce02e2a2f515d0092368f17560b4314a662cc6b0ae3bd
+size 1003497024

XL-LuaCopilot-1.7B-FFT-Q3_K_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d06ab461534367d70e6161938ae3e173e05db61ca3ec19e048743def920aa1a
+size 939533888

XL-LuaCopilot-1.7B-FFT-Q3_K_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea5448930564e5a6f3daca95ac2a28fab61dd77db8d72f019ebd511eaa5cc556
+size 867247680

XL-LuaCopilot-1.7B-FFT-Q4_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:722bb4e698ce94eca1c7bd4b6ed9b5d4562426783e5336c24a65cc46f292cae1
+size 1054418496

XL-LuaCopilot-1.7B-FFT-Q4_1.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1a32b5088a56f37ac8af7db6f1ad7f393e3949dfd717e4943ccefb4a5de36c5
+size 1142498880

XL-LuaCopilot-1.7B-FFT-Q4_K.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7379ec6d98eb94408d63db3cc5ef5d7b9d7f63e380177a3ca6d8259184d8d298
+size 1107404352

XL-LuaCopilot-1.7B-FFT-Q4_K_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7379ec6d98eb94408d63db3cc5ef5d7b9d7f63e380177a3ca6d8259184d8d298
+size 1107404352

XL-LuaCopilot-1.7B-FFT-Q4_K_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9198790fa8d492990467f615ce906e8e1ad5fe881c4d951e4f84c0b715f94716
+size 1060185664

XL-LuaCopilot-1.7B-FFT-Q5_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a818d8d6f0f798c18638a2390c77663e3c6c9db6ecb0171c54ebb265523bf617
+size 1230579264

XL-LuaCopilot-1.7B-FFT-Q5_1.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b57452e2ba7043e6299248c134a206907eb4f0107a2ad2405b8d1dbd3d6b24ec
+size 1318659648

XL-LuaCopilot-1.7B-FFT-Q5_K.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5d5735b9575c18cfcc2ed851ac6dd76f82e88704d852020f238ac390578fa8c
+size 1257875008

XL-LuaCopilot-1.7B-FFT-Q5_K_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5d5735b9575c18cfcc2ed851ac6dd76f82e88704d852020f238ac390578fa8c
+size 1257875008

XL-LuaCopilot-1.7B-FFT-Q5_K_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4707bd98d25f78bc5f9cece5ae3a296dda6588d10da57026a8ed43b298130f5e
+size 1230579264

XL-LuaCopilot-1.7B-FFT-Q6_K.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46f5ef40d7083057500dcb66c2b454957bf2e1d83f25e431331b02d2538118ef
+size 1417750080

XL-LuaCopilot-1.7B-FFT-Q8_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de830da67aee50ae76268213b3e00375c8cde6180d1b0b089cb26e67e50715dd
+size 1834421824