first commit.

README.md

@@ -1,28 +1,76 @@
 ---
-language:
-- ja
-base_model:
-- elyza/ELYZA-japanese-Llama-2-13b-instruct
+base_model: elyza/ELYZA-japanese-Llama-2-7b-instruct
 tags:
-- medusa
+- generated_from_trainer
+model-index:
+- name: medusa-ELYZA-japanese-Llama-2-7b-instruct
+  results: []
 ---
 
-# outputs
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
+# Elyza-japanese-Llama-2-7b-instruct_qlora_ja_conv_stage2
+
+This model is a fine-tuned version of [elyza/ELYZA-japanese-Llama-2-7b-instruct](https://huggingface.co/elyza/ELYZA-japanese-Llama-2-7b-instruct) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.3897
+
+## Model description
 
 This is a Medusa-2 created using [Medusa](https://github.com/FasterDecoding/Medusa).
- 
 
-## Details
+## Intended uses & limitations
+
+- [$B!Z(BOrion-14B Series$B![(B Models Community License Agreement](https://huggingface.co/OrionStarAI/Orion-14B-Chat/blob/main/ModelsCommunityLicenseAgreement)
+
+## Training and evaluation data
+
+- [shi3z/ja_conv_wikipedia_orion14B_100K](https://huggingface.co/datasets/shi3z/ja_conv_wikipedia_orion14B_100K)
+
+## Training procedure
 
-### Base Model
+### Training hyperparameters
 
-* [elyza/ELYZA-japanese-Llama-2-7b-instruct](https://huggingface.co/elyza/ELYZA-japanese-Llama-2-7b-instruct)
+The following hyperparameters were used during training:
+- learning_rate: 0.0005
+- train_batch_size: 4
+- eval_batch_size: 4
+- seed: 42
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 16
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 40
+- num_epochs: 1
 
-### Training dataset
+### Training results
 
-* [shi3z/ja_conv_wikipedia_orion14B_100K](https://huggingface.co/datasets/shi3z/ja_conv_wikipedia_orion14B_100K)
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 2.7245        | 0.06  | 40   | 2.7628          |
+| 2.5599        | 0.11  | 80   | 2.6823          |
+| 2.5001        | 0.17  | 120  | 2.6359          |
+| 2.5642        | 0.23  | 160  | 2.6046          |
+| 2.4892        | 0.28  | 200  | 2.5834          |
+| 2.5232        | 0.34  | 240  | 2.5551          |
+| 2.463         | 0.4   | 280  | 2.5327          |
+| 2.3983        | 0.45  | 320  | 2.5097          |
+| 2.4192        | 0.51  | 360  | 2.4880          |
+| 2.4186        | 0.57  | 400  | 2.4677          |
+| 2.4301        | 0.62  | 440  | 2.4465          |
+| 2.4194        | 0.68  | 480  | 2.4293          |
+| 2.392         | 0.74  | 520  | 2.4144          |
+| 2.4235        | 0.79  | 560  | 2.4014          |
+| 2.4328        | 0.85  | 600  | 2.3939          |
+| 2.456         | 0.91  | 640  | 2.3906          |
+| 2.4135        | 0.96  | 680  | 2.3897          |
 
 
-# License
+### Framework versions
 
-* [【Orion-14B Series】 Models Community License Agreement](https://huggingface.co/OrionStarAI/Orion-14B-Chat/blob/main/ModelsCommunityLicenseAgreement)
+- Transformers 4.34.1
+- Pytorch 2.1.2
+- Datasets 2.16.1
+- Tokenizers 0.14.1

adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "elyza/ELYZA-japanese-Llama-2-7b-instruct",
+  "bias": "none",
+  "fan_in_fan_out": null,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": [
+    "medusa_head.0",
+    "medusa_head.1",
+    "medusa_head.2",
+    "medusa_head.3",
+    "medusa_head.4"
+  ],
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "v_proj",
+    "o_proj",
+    "down_proj",
+    "lm_head",
+    "k_proj",
+    "up_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_rslora": false
+}

adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:499f3ca7e20d11ae3d4718e549f30f444d9fa9e2ad0b4d5a967895b56f6aed2c
+size 2062971154

config.json

@@ -0,0 +1,42 @@
+{
+  "_name_or_path": "elyza/ELYZA-japanese-Llama-2-7b-instruct",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "medusa_num_heads": 5,
+  "medusa_num_layers": 1,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pad_token_id": 0,
+  "pretraining_tp": 1,
+  "quantization_config": {
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.34.1",
+  "use_cache": false,
+  "vocab_size": 32000
+}

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

tokenizer_config.json

@@ -0,0 +1,44 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "trust_remote_code": false,
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true,
+  "use_fast": true
+}