初始化项目，由ModelHub XC社区提供模型

Model: shisa-ai/ablation-73-rafathenev2.rp.tl.unphi42e6-shisa-v2-unphi-4-14b Source: Original Platform
2026-05-01 20:17:26 +08:00
commit 516ec77149
19 changed files with 101516 additions and 0 deletions
--- a/README.md
+++ b/README.md
@@ -0,0 +1,198 @@
+---
+library_name: transformers
+license: mit
+base_model: unsloth/phi-4
+tags:
+- generated_from_trainer
+datasets:
+- shisa-ai/shisa-v1-athenev2-reannotated-filtered
+- shisa-ai/shisa-v2-roleplaying
+- shisa-ai/translation-master-set
+model-index:
+- name: outputs/ablation-73-rafathenev2.rp.tl.unphi42e6-shisa-v2-unphi-4-14b
+  results: []
+---
+
+<!-- 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/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
+<details><summary>See axolotl config</summary>
+
+axolotl version: `0.8.0.dev0`
+```yaml
+# 33 but w/ 5e-6 LR
+
+base_model: unsloth/phi-4
+model_type: AutoModelForCausalLM
+tokenizer_type: AutoTokenizer
+
+load_in_8bit: false
+load_in_4bit: false
+strict: false
+
+# User Liger
+plugins:
+  - axolotl.integrations.liger.LigerPlugin
+liger_rope: true
+liger_rms_norm: true
+liger_glu_activation: true
+liger_fused_linear_cross_entropy: true
+
+chat_template: tokenizer_default
+datasets:
+  - path: shisa-ai/shisa-v1-athenev2-reannotated-filtered
+    type: chat_template
+    field_messages: conversations
+    message_field_role: from
+    message_field_content: value
+  - path: shisa-ai/shisa-v2-roleplaying
+    type: chat_template
+    field_messages: conversations
+    message_property_mappings:
+      role: role
+      content: content
+    roles:
+      system:
+        - system
+      assistant:
+        - gpt
+        - model
+        - assistant
+      user:
+        - human
+        - user
+    roles_to_train: ["assistant"]
+  - path: shisa-ai/translation-master-set
+    type: chat_template
+    field_messages: conversations
+    message_property_mappings:
+      role: role
+      content: content
+    roles:
+      system:
+        - system
+      assistant:
+        - gpt
+        - model
+        - assistant
+      user:
+        - human
+        - user
+    roles_to_train: ["assistant"]
+dataset_prepared_path: last_run_prepared
+val_set_size: 0.05
+output_dir: ./outputs/ablation-73-rafathenev2.rp.tl.unphi42e6-shisa-v2-unphi-4-14b
+
+sequence_len: 8192
+sample_packing: true
+pad_to_sequence_len: true
+
+# marginal difference
+neftune_noise_alpha: 5
+
+use_wandb: true
+wandb_project: shisa-v2
+wandb_entity: augmxnt
+wandb_name: ablation-73-rafathenev2.rp.tl.unphi42e6-shisa-v2-unphi-4-14b
+
+gradient_accumulation_steps: 2
+micro_batch_size: 4
+num_epochs: 3
+optimizer: paged_adamw_8bit
+lr_scheduler: linear
+learning_rate: 2e-6
+
+train_on_inputs: false
+group_by_length: false
+bf16: auto
+fp16:
+tf32: false
+
+gradient_checkpointing: true
+gradient_checkpointing_kwargs:
+  use_reentrant: false
+early_stopping_patience:
+resume_from_checkpoint:
+logging_steps: 1
+xformers_attention:
+flash_attention: true
+
+warmup_ratio: 0.1
+evals_per_epoch: 4
+eval_table_size:
+saves_per_epoch: 0
+save_total_limit: 1 # Only store a single checkpoint
+debug:
+deepspeed: zero3_bf16.json
+weight_decay: 0.0001
+fsdp:
+fsdp_config:
+special_tokens:
+  pad_token: <|end_of_text|>
+
+```
+
+</details><br>
+
+# outputs/ablation-73-rafathenev2.rp.tl.unphi42e6-shisa-v2-unphi-4-14b
+
+This model is a fine-tuned version of [unsloth/phi-4](https://huggingface.co/unsloth/phi-4) on the shisa-ai/shisa-v1-athenev2-reannotated-filtered, the shisa-ai/shisa-v2-roleplaying and the shisa-ai/translation-master-set datasets.
+It achieves the following results on the evaluation set:
+- Loss: 0.4463
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 2e-06
+- train_batch_size: 4
+- eval_batch_size: 4
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 64
+- total_eval_batch_size: 32
+- optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_steps: 175
+- num_epochs: 3.0
+
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.6323        | 0.0017 | 1    | 0.6161          |
+| 0.524         | 0.2506 | 147  | 0.5063          |
+| 0.4596        | 0.5013 | 294  | 0.4791          |
+| 0.4552        | 0.7519 | 441  | 0.4675          |
+| 0.4389        | 1.0017 | 588  | 0.4603          |
+| 0.4312        | 1.2523 | 735  | 0.4562          |
+| 0.4404        | 1.5030 | 882  | 0.4527          |
+| 0.4712        | 1.7536 | 1029 | 0.4498          |
+| 0.4035        | 2.0034 | 1176 | 0.4477          |
+| 0.4018        | 2.2540 | 1323 | 0.4480          |
+| 0.3889        | 2.5047 | 1470 | 0.4469          |
+| 0.4117        | 2.7553 | 1617 | 0.4463          |
+
+
+### Framework versions
+
+- Transformers 4.49.0
+- Pytorch 2.6.0+cu124
+- Datasets 3.2.0
+- Tokenizers 0.21.0