Upload folder using huggingface_hub

README.md

@@ -0,0 +1,57 @@
+---
+base_model: deepseek-ai/DeepSeek-V2-Lite
+datasets: RoxanneWsyw/ESFT-summary
+library_name: transformers
+tags:
+- generated_from_trainer
+- open-r1
+licence: license
+---
+
+# Model Card for None
+
+This model is a fine-tuned version of [deepseek-ai/DeepSeek-V2-Lite](https://huggingface.co/deepseek-ai/DeepSeek-V2-Lite) on the [RoxanneWsyw/ESFT-summary](https://huggingface.co/datasets/RoxanneWsyw/ESFT-summary) dataset.
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/jayzxinkai-uc-san-diego/moe-honing/runs/h40v6ilw) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.16.0.dev0
+- Transformers: 4.49.0
+- Pytorch: 2.6.0
+- Datasets: 4.8.3
+- Tokenizers: 0.21.4
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

all_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 1.6538790970985021e+18,
+    "train_loss": 1.2438032255853926,
+    "train_runtime": 8525.9707,
+    "train_samples": 19587,
+    "train_samples_per_second": 2.297,
+    "train_steps_per_second": 0.144
+}

config.json

@@ -0,0 +1,61 @@
+{
+  "_name_or_path": "deepseek-ai/DeepSeek-V2-Lite",
+  "architectures": [
+    "DeepseekV2ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoConfig": "deepseek-ai/DeepSeek-V2-Lite--configuration_deepseek.DeepseekV2Config",
+    "AutoModel": "deepseek-ai/DeepSeek-V2-Lite--modeling_deepseek.DeepseekV2Model",
+    "AutoModelForCausalLM": "deepseek-ai/DeepSeek-V2-Lite--modeling_deepseek.DeepseekV2ForCausalLM"
+  },
+  "aux_loss_alpha": 0.001,
+  "bos_token_id": 100000,
+  "eos_token_id": 100001,
+  "ep_size": 1,
+  "first_k_dense_replace": 1,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 10944,
+  "kv_lora_rank": 512,
+  "max_position_embeddings": 163840,
+  "model_type": "deepseek_v2",
+  "moe_intermediate_size": 1408,
+  "moe_layer_freq": 1,
+  "n_group": 1,
+  "n_routed_experts": 64,
+  "n_shared_experts": 2,
+  "norm_topk_prob": false,
+  "num_attention_heads": 16,
+  "num_experts_per_tok": 6,
+  "num_hidden_layers": 27,
+  "num_key_value_heads": 16,
+  "pretraining_tp": 1,
+  "q_lora_rank": null,
+  "qk_nope_head_dim": 128,
+  "qk_rope_head_dim": 64,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "beta_fast": 32,
+    "beta_slow": 1,
+    "factor": 40,
+    "mscale": 0.707,
+    "mscale_all_dim": 0.707,
+    "original_max_position_embeddings": 4096,
+    "type": "yarn"
+  },
+  "rope_theta": 10000,
+  "routed_scaling_factor": 1.0,
+  "scoring_func": "softmax",
+  "seq_aux": true,
+  "tie_word_embeddings": false,
+  "topk_group": 1,
+  "topk_method": "greedy",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "v_head_dim": 128,
+  "vocab_size": 102400
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 100000,
+  "do_sample": true,
+  "eos_token_id": 100001,
+  "temperature": 0.3,
+  "top_p": 0.95,
+  "transformers_version": "4.49.0"
+}

model-00001-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2457ec9edd72f8351039664d1f3272d618e274e32c3c573adec6003d96feed98
+size 4994763632

model-00002-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a19ebf28284f39d2c5f2768fa231635834e1662c859b8787a2aa67f17a6a193c
+size 4995044944

model-00003-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d3541a3afdf6e54c41d1e64c12cb9e278d6306b3ce2d7332ef30e6cbef9f1b4
+size 4996085000

model-00004-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eadfb0e42fcd21ed166f5f016e4ea87d71d3ae1398e1858592b39e1cc05b5914
+size 4996085224

model-00005-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d8ad850f9127800df6c2a1146898b17704a9fb17f3432b84e88a42cb91f533c
+size 4996085224

model-00006-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09d36d1ff84020644b29166297da9f378e304fbb4541ca0f13ba9857a809ea61
+size 4995045792

model-00007-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75d551b9802edc5ab1cb0f8a3d1ce0dbece82366e4b1495ba175666a8f778586
+size 1440515736

special_tokens_map.json

@@ -0,0 +1,17 @@
+{
+  "bos_token": {
+    "content": "<｜begin▁of▁sentence｜>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<｜end▁of▁sentence｜>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<｜end▁of▁sentence｜>"
+}

tokenizer_config.json

@@ -0,0 +1,36 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "100000": {
+      "content": "<｜begin▁of▁sentence｜>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100001": {
+      "content": "<｜end▁of▁sentence｜>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<｜begin▁of▁sentence｜>",
+  "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ 'User: ' + message['content'] + '\n\n' }}{% elif message['role'] == 'assistant' %}{{ 'Assistant: ' + message['content'] + eos_token }}{% elif message['role'] == 'system' %}{{ message['content'] + '\n\n' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<｜end▁of▁sentence｜>",
+  "extra_special_tokens": {},
+  "fast_tokenizer": true,
+  "legacy": true,
+  "model_max_length": 16384,
+  "pad_token": "<｜end▁of▁sentence｜>",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizerFast",
+  "unk_token": null,
+  "use_default_system_prompt": false
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 1.6538790970985021e+18,
+    "train_loss": 1.2438032255853926,
+    "train_runtime": 8525.9707,
+    "train_samples": 19587,
+    "train_samples_per_second": 2.297,
+    "train_steps_per_second": 0.144
+}

training.log

@@ -0,0 +1,737 @@
+2026-03-21 20:09:11 - INFO - __main__ - Model parameters ModelConfig(model_name_or_path='deepseek-ai/DeepSeek-V2-Lite', model_revision='main', torch_dtype='bfloat16', trust_remote_code=True, attn_implementation='flash_attention_2', use_peft=False, lora_r=16, lora_alpha=32, lora_dropout=0.05, lora_target_modules=None, lora_modules_to_save=None, lora_task_type='CAUSAL_LM', use_rslora=False, load_in_8bit=False, load_in_4bit=False, bnb_4bit_quant_type='nf4', use_bnb_nested_quant=False)
+2026-03-21 20:09:11 - INFO - __main__ - Script parameters ScriptArguments(dataset_name='RoxanneWsyw/ESFT-summary', dataset_config=None, dataset_train_split='train', dataset_test_split='test', gradient_checkpointing_use_reentrant=False, ignore_bias_buffers=False)
+2026-03-21 20:09:11 - INFO - __main__ - Training parameters SFTConfig(
+_n_gpu=1,
+accelerator_config={'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False},
+adafactor=False,
+adam_beta1=0.9,
+adam_beta2=0.999,
+adam_epsilon=1e-08,
+attn_kl_weight=1.0,
+auto_find_batch_size=False,
+average_tokens_across_devices=False,
+batch_eval_metrics=False,
+benchmarks=[],
+bf16=True,
+bf16_full_eval=False,
+callbacks=[],
+chars_per_token=<CHARS_PER_TOKEN>,
+chat_template=None,
+cluster_mode=hierarchical-dynamic,
+cluster_num_groups=None,
+cluster_prune_ratio=None,
+cluster_prune_tau=1.0,
+data_seed=None,
+dataloader_drop_last=False,
+dataloader_num_workers=0,
+dataloader_persistent_workers=False,
+dataloader_pin_memory=True,
+dataloader_prefetch_factor=None,
+dataset_batch_size=None,
+dataset_kwargs=None,
+dataset_num_proc=None,
+dataset_text_field=text,
+ddp_backend=None,
+ddp_broadcast_buffers=None,
+ddp_bucket_cap_mb=None,
+ddp_find_unused_parameters=None,
+ddp_timeout=1800000000,
+debug=[],
+deepspeed=None,
+disable_teacher_dropout=True,
+disable_tqdm=False,
+dispatch_batches=None,
+do_eval=True,
+do_predict=False,
+do_train=False,
+entropy_slope_alpha=1.0,
+entropy_slope_beta=1.0,
+eval_accumulation_steps=None,
+eval_delay=0,
+eval_do_concat_batches=True,
+eval_on_start=False,
+eval_packing=None,
+eval_steps=None,
+eval_strategy=IntervalStrategy.NO,
+eval_use_gather_object=False,
+evaluation_strategy=None,
+fp16=False,
+fp16_backend=auto,
+fp16_full_eval=False,
+fp16_opt_level=O1,
+fsdp=[],
+fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False},
+fsdp_min_num_params=0,
+fsdp_transformer_layer_cls_to_wrap=None,
+full_determinism=False,
+gradient_accumulation_steps=1,
+gradient_checkpointing=True,
+gradient_checkpointing_kwargs={'use_reentrant': False},
+greater_is_better=None,
+group_by_length=False,
+half_precision_backend=auto,
+hub_always_push=False,
+hub_model_id=None,
+hub_model_revision=main,
+hub_private_repo=None,
+hub_strategy=HubStrategy.EVERY_SAVE,
+hub_token=<HUB_TOKEN>,
+ignore_data_skip=False,
+include_for_metrics=[],
+include_inputs_for_metrics=False,
+include_num_input_tokens_seen=False,
+include_tokens_per_second=False,
+jit_mode_eval=False,
+label_names=None,
+label_smoothing_factor=0.0,
+last_entropy_weight=1.0,
+layer_entropy_l1_layers=None,
+layer_entropy_l1_weight=1.0,
+learning_rate=1e-05,
+length_column_name=length,
+load_best_model_at_end=False,
+local_rank=0,
+log_level=info,
+log_level_replica=warning,
+log_on_each_node=True,
+logging_dir=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft/runs/Mar21_20-09-09_orchard-community-1,
+logging_first_step=False,
+logging_nan_inf_filter=True,
+logging_steps=1,
+logging_strategy=IntervalStrategy.STEPS,
+lr_scheduler_kwargs={'min_lr_rate': 0.1},
+lr_scheduler_type=SchedulerType.COSINE_WITH_MIN_LR,
+max_grad_norm=1.0,
+max_length=4096,
+max_seq_length=None,
+max_steps=-1,
+merging_metrics=None,
+metric_for_best_model=None,
+model_init_kwargs=None,
+mp_parameters=,
+neftune_noise_alpha=None,
+no_cuda=False,
+num_of_sequences=None,
+num_train_epochs=1,
+optim=OptimizerNames.ADAMW_TORCH,
+optim_args=None,
+optim_target_modules=None,
+output_dir=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft,
+overwrite_hub_revision=False,
+overwrite_output_dir=True,
+packing=False,
+past_index=-1,
+per_device_eval_batch_size=16,
+per_device_train_batch_size=4,
+prediction_loss_only=False,
+push_to_hub=False,
+push_to_hub_model_id=None,
+push_to_hub_organization=None,
+push_to_hub_revision=False,
+push_to_hub_token=<PUSH_TO_HUB_TOKEN>,
+ray_scope=last,
+remove_unused_columns=True,
+report_to=['wandb'],
+restore_callback_states_from_checkpoint=False,
+resume_from_checkpoint=None,
+router_manual_mask=None,
+router_prune_enable=True,
+router_prune_expert_per_layer=None,
+router_prune_interval=5,
+router_prune_min_keep=1,
+router_prune_start_step=None,
+router_prune_step_size=32,
+router_prune_use_plan=True,
+run_name=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft,
+save_on_each_node=False,
+save_only_model=False,
+save_safetensors=True,
+save_steps=500,
+save_strategy=SaveStrategy.NO,
+save_total_limit=None,
+seed=1234,
+skip_memory_metrics=True,
+split_batches=None,
+system_prompt=None,
+teacher_attn_implementation=None,
+teacher_model_name_or_path=None,
+teacher_model_revision=None,
+teacher_torch_dtype=auto,
+tf32=None,
+torch_compile=False,
+torch_compile_backend=None,
+torch_compile_mode=None,
+torch_empty_cache_steps=None,
+torchdynamo=None,
+tpu_metrics_debug=False,
+tpu_num_cores=None,
+use_cpu=False,
+use_ipex=False,
+use_legacy_prediction_loop=False,
+use_liger=False,
+use_liger_kernel=False,
+use_mps_device=False,
+wandb_entity=jayzxinkai-uc-san-diego,
+wandb_project=moe-honing,
+warmup_ratio=0.1,
+warmup_steps=0,
+weight_decay=0.0,
+weight_feature_rank=None,
+)
+2026-03-21 20:09:12 - INFO - datasets.builder - Found cached dataset esft-summary (/tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138)
+2026-03-21 20:09:12 - INFO - datasets.arrow_dataset - Caching processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-d8c6f402c91a2432.arrow
+2026-03-21 20:09:13 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-1975b68b541b14ab_*_of_00001.arrow
+2026-03-21 20:09:14 - INFO - __main__ - *** Initializing model kwargs ***
+2026-03-21 20:10:05 - INFO - datasets.arrow_dataset - Caching processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-6435616c9a34cd0e.arrow
+2026-03-21 20:10:06 - INFO - datasets.arrow_dataset - Caching processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-264b8deb588933bd.arrow
+2026-03-21 20:10:07 - INFO - datasets.arrow_dataset - Caching processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-ad7de21f8027b38f.arrow
+2026-03-21 20:10:34 - INFO - datasets.arrow_dataset - Caching processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-545571e95ce7027c.arrow
+2026-03-21 20:10:47 - INFO - __main__ - *** Train ***
+2026-03-21 20:10:47 - INFO - __main__ - DeepseekV2ForCausalLM(
+  (model): DeepseekV2Model(
+    (embed_tokens): Embedding(102400, 2048)
+    (layers): ModuleList(
+      (0): DeepseekV2DecoderLayer(
+        (self_attn): DeepseekV2FlashAttention2(
+          (q_proj): Linear(in_features=2048, out_features=3072, bias=False)
+          (kv_a_proj_with_mqa): Linear(in_features=2048, out_features=576, bias=False)
+          (kv_a_layernorm): DeepseekV2RMSNorm()
+          (kv_b_proj): Linear(in_features=512, out_features=4096, bias=False)
+          (o_proj): Linear(in_features=2048, out_features=2048, bias=False)
+          (rotary_emb): DeepseekV2YarnRotaryEmbedding()
+        )
+        (mlp): DeepseekV2MLP(
+          (gate_proj): Linear(in_features=2048, out_features=10944, bias=False)
+          (up_proj): Linear(in_features=2048, out_features=10944, bias=False)
+          (down_proj): Linear(in_features=10944, out_features=2048, bias=False)
+          (act_fn): SiLU()
+        )
+        (input_layernorm): DeepseekV2RMSNorm()
+        (post_attention_layernorm): DeepseekV2RMSNorm()
+      )
+      (1-26): 26 x DeepseekV2DecoderLayer(
+        (self_attn): DeepseekV2FlashAttention2(
+          (q_proj): Linear(in_features=2048, out_features=3072, bias=False)
+          (kv_a_proj_with_mqa): Linear(in_features=2048, out_features=576, bias=False)
+          (kv_a_layernorm): DeepseekV2RMSNorm()
+          (kv_b_proj): Linear(in_features=512, out_features=4096, bias=False)
+          (o_proj): Linear(in_features=2048, out_features=2048, bias=False)
+          (rotary_emb): DeepseekV2YarnRotaryEmbedding()
+        )
+        (mlp): DeepseekV2MoE(
+          (experts): ModuleList(
+            (0-63): 64 x DeepseekV2MLP(
+              (gate_proj): Linear(in_features=2048, out_features=1408, bias=False)
+              (up_proj): Linear(in_features=2048, out_features=1408, bias=False)
+              (down_proj): Linear(in_features=1408, out_features=2048, bias=False)
+              (act_fn): SiLU()
+            )
+          )
+          (gate): MoEGate()
+          (shared_experts): DeepseekV2MLP(
+            (gate_proj): Linear(in_features=2048, out_features=2816, bias=False)
+            (up_proj): Linear(in_features=2048, out_features=2816, bias=False)
+            (down_proj): Linear(in_features=2816, out_features=2048, bias=False)
+            (act_fn): SiLU()
+          )
+        )
+        (input_layernorm): DeepseekV2RMSNorm()
+        (post_attention_layernorm): DeepseekV2RMSNorm()
+      )
+    )
+    (norm): DeepseekV2RMSNorm()
+  )
+  (lm_head): Linear(in_features=2048, out_features=102400, bias=False)
+)
+2026-03-21 20:13:27 - INFO - __main__ - Model parameters ModelConfig(model_name_or_path='deepseek-ai/DeepSeek-V2-Lite', model_revision='main', torch_dtype='bfloat16', trust_remote_code=True, attn_implementation='flash_attention_2', use_peft=False, lora_r=16, lora_alpha=32, lora_dropout=0.05, lora_target_modules=None, lora_modules_to_save=None, lora_task_type='CAUSAL_LM', use_rslora=False, load_in_8bit=False, load_in_4bit=False, bnb_4bit_quant_type='nf4', use_bnb_nested_quant=False)
+2026-03-21 20:13:27 - INFO - __main__ - Script parameters ScriptArguments(dataset_name='RoxanneWsyw/ESFT-summary', dataset_config=None, dataset_train_split='train', dataset_test_split='test', gradient_checkpointing_use_reentrant=False, ignore_bias_buffers=False)
+2026-03-21 20:13:27 - INFO - __main__ - Training parameters SFTConfig(
+_n_gpu=1,
+accelerator_config={'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False},
+adafactor=False,
+adam_beta1=0.9,
+adam_beta2=0.999,
+adam_epsilon=1e-08,
+attn_kl_weight=1.0,
+auto_find_batch_size=False,
+average_tokens_across_devices=False,
+batch_eval_metrics=False,
+benchmarks=[],
+bf16=True,
+bf16_full_eval=False,
+callbacks=[],
+chars_per_token=<CHARS_PER_TOKEN>,
+chat_template=None,
+cluster_mode=hierarchical-dynamic,
+cluster_num_groups=None,
+cluster_prune_ratio=None,
+cluster_prune_tau=1.0,
+data_seed=None,
+dataloader_drop_last=False,
+dataloader_num_workers=0,
+dataloader_persistent_workers=False,
+dataloader_pin_memory=True,
+dataloader_prefetch_factor=None,
+dataset_batch_size=None,
+dataset_kwargs=None,
+dataset_num_proc=None,
+dataset_text_field=text,
+ddp_backend=None,
+ddp_broadcast_buffers=None,
+ddp_bucket_cap_mb=None,
+ddp_find_unused_parameters=None,
+ddp_timeout=1800000000,
+debug=[],
+deepspeed=None,
+disable_teacher_dropout=True,
+disable_tqdm=False,
+dispatch_batches=None,
+do_eval=True,
+do_predict=False,
+do_train=False,
+entropy_slope_alpha=1.0,
+entropy_slope_beta=1.0,
+eval_accumulation_steps=None,
+eval_delay=0,
+eval_do_concat_batches=True,
+eval_on_start=False,
+eval_packing=None,
+eval_steps=None,
+eval_strategy=IntervalStrategy.NO,
+eval_use_gather_object=False,
+evaluation_strategy=None,
+fp16=False,
+fp16_backend=auto,
+fp16_full_eval=False,
+fp16_opt_level=O1,
+fsdp=[],
+fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False},
+fsdp_min_num_params=0,
+fsdp_transformer_layer_cls_to_wrap=None,
+full_determinism=False,
+gradient_accumulation_steps=1,
+gradient_checkpointing=True,
+gradient_checkpointing_kwargs={'use_reentrant': False},
+greater_is_better=None,
+group_by_length=False,
+half_precision_backend=auto,
+hub_always_push=False,
+hub_model_id=None,
+hub_model_revision=main,
+hub_private_repo=None,
+hub_strategy=HubStrategy.EVERY_SAVE,
+hub_token=<HUB_TOKEN>,
+ignore_data_skip=False,
+include_for_metrics=[],
+include_inputs_for_metrics=False,
+include_num_input_tokens_seen=False,
+include_tokens_per_second=False,
+jit_mode_eval=False,
+label_names=None,
+label_smoothing_factor=0.0,
+last_entropy_weight=1.0,
+layer_entropy_l1_layers=None,
+layer_entropy_l1_weight=1.0,
+learning_rate=1e-05,
+length_column_name=length,
+load_best_model_at_end=False,
+local_rank=0,
+log_level=info,
+log_level_replica=warning,
+log_on_each_node=True,
+logging_dir=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft/runs/Mar21_20-13-25_orchard-community-1,
+logging_first_step=False,
+logging_nan_inf_filter=True,
+logging_steps=1,
+logging_strategy=IntervalStrategy.STEPS,
+lr_scheduler_kwargs={'min_lr_rate': 0.1},
+lr_scheduler_type=SchedulerType.COSINE_WITH_MIN_LR,
+max_grad_norm=1.0,
+max_length=4096,
+max_seq_length=None,
+max_steps=-1,
+merging_metrics=None,
+metric_for_best_model=None,
+model_init_kwargs=None,
+mp_parameters=,
+neftune_noise_alpha=None,
+no_cuda=False,
+num_of_sequences=None,
+num_train_epochs=1,
+optim=OptimizerNames.ADAMW_TORCH,
+optim_args=None,
+optim_target_modules=None,
+output_dir=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft,
+overwrite_hub_revision=False,
+overwrite_output_dir=True,
+packing=False,
+past_index=-1,
+per_device_eval_batch_size=16,
+per_device_train_batch_size=2,
+prediction_loss_only=False,
+push_to_hub=False,
+push_to_hub_model_id=None,
+push_to_hub_organization=None,
+push_to_hub_revision=False,
+push_to_hub_token=<PUSH_TO_HUB_TOKEN>,
+ray_scope=last,
+remove_unused_columns=True,
+report_to=['wandb'],
+restore_callback_states_from_checkpoint=False,
+resume_from_checkpoint=None,
+router_manual_mask=None,
+router_prune_enable=True,
+router_prune_expert_per_layer=None,
+router_prune_interval=5,
+router_prune_min_keep=1,
+router_prune_start_step=None,
+router_prune_step_size=32,
+router_prune_use_plan=True,
+run_name=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft,
+save_on_each_node=False,
+save_only_model=False,
+save_safetensors=True,
+save_steps=500,
+save_strategy=SaveStrategy.NO,
+save_total_limit=None,
+seed=1234,
+skip_memory_metrics=True,
+split_batches=None,
+system_prompt=None,
+teacher_attn_implementation=None,
+teacher_model_name_or_path=None,
+teacher_model_revision=None,
+teacher_torch_dtype=auto,
+tf32=None,
+torch_compile=False,
+torch_compile_backend=None,
+torch_compile_mode=None,
+torch_empty_cache_steps=None,
+torchdynamo=None,
+tpu_metrics_debug=False,
+tpu_num_cores=None,
+use_cpu=False,
+use_ipex=False,
+use_legacy_prediction_loop=False,
+use_liger=False,
+use_liger_kernel=False,
+use_mps_device=False,
+wandb_entity=jayzxinkai-uc-san-diego,
+wandb_project=moe-honing,
+warmup_ratio=0.1,
+warmup_steps=0,
+weight_decay=0.0,
+weight_feature_rank=None,
+)
+2026-03-21 20:13:28 - INFO - datasets.builder - Found cached dataset esft-summary (/tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138)
+2026-03-21 20:13:28 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-d8c6f402c91a2432_*_of_00001.arrow
+2026-03-21 20:13:28 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-1975b68b541b14ab_*_of_00001.arrow
+2026-03-21 20:13:29 - INFO - __main__ - *** Initializing model kwargs ***
+2026-03-21 20:13:47 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-6435616c9a34cd0e_*_of_00001.arrow
+2026-03-21 20:13:47 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-264b8deb588933bd_*_of_00001.arrow
+2026-03-21 20:13:47 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-ad7de21f8027b38f_*_of_00001.arrow
+2026-03-21 20:13:47 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-545571e95ce7027c_*_of_00001.arrow
+2026-03-21 20:13:53 - INFO - __main__ - *** Train ***
+2026-03-21 20:13:53 - INFO - __main__ - DeepseekV2ForCausalLM(
+  (model): DeepseekV2Model(
+    (embed_tokens): Embedding(102400, 2048)
+    (layers): ModuleList(
+      (0): DeepseekV2DecoderLayer(
+        (self_attn): DeepseekV2FlashAttention2(
+          (q_proj): Linear(in_features=2048, out_features=3072, bias=False)
+          (kv_a_proj_with_mqa): Linear(in_features=2048, out_features=576, bias=False)
+          (kv_a_layernorm): DeepseekV2RMSNorm()
+          (kv_b_proj): Linear(in_features=512, out_features=4096, bias=False)
+          (o_proj): Linear(in_features=2048, out_features=2048, bias=False)
+          (rotary_emb): DeepseekV2YarnRotaryEmbedding()
+        )
+        (mlp): DeepseekV2MLP(
+          (gate_proj): Linear(in_features=2048, out_features=10944, bias=False)
+          (up_proj): Linear(in_features=2048, out_features=10944, bias=False)
+          (down_proj): Linear(in_features=10944, out_features=2048, bias=False)
+          (act_fn): SiLU()
+        )
+        (input_layernorm): DeepseekV2RMSNorm()
+        (post_attention_layernorm): DeepseekV2RMSNorm()
+      )
+      (1-26): 26 x DeepseekV2DecoderLayer(
+        (self_attn): DeepseekV2FlashAttention2(
+          (q_proj): Linear(in_features=2048, out_features=3072, bias=False)
+          (kv_a_proj_with_mqa): Linear(in_features=2048, out_features=576, bias=False)
+          (kv_a_layernorm): DeepseekV2RMSNorm()
+          (kv_b_proj): Linear(in_features=512, out_features=4096, bias=False)
+          (o_proj): Linear(in_features=2048, out_features=2048, bias=False)
+          (rotary_emb): DeepseekV2YarnRotaryEmbedding()
+        )
+        (mlp): DeepseekV2MoE(
+          (experts): ModuleList(
+            (0-63): 64 x DeepseekV2MLP(
+              (gate_proj): Linear(in_features=2048, out_features=1408, bias=False)
+              (up_proj): Linear(in_features=2048, out_features=1408, bias=False)
+              (down_proj): Linear(in_features=1408, out_features=2048, bias=False)
+              (act_fn): SiLU()
+            )
+          )
+          (gate): MoEGate()
+          (shared_experts): DeepseekV2MLP(
+            (gate_proj): Linear(in_features=2048, out_features=2816, bias=False)
+            (up_proj): Linear(in_features=2048, out_features=2816, bias=False)
+            (down_proj): Linear(in_features=2816, out_features=2048, bias=False)
+            (act_fn): SiLU()
+          )
+        )
+        (input_layernorm): DeepseekV2RMSNorm()
+        (post_attention_layernorm): DeepseekV2RMSNorm()
+      )
+    )
+    (norm): DeepseekV2RMSNorm()
+  )
+  (lm_head): Linear(in_features=2048, out_features=102400, bias=False)
+)
+2026-03-21 20:15:46 - INFO - __main__ - Model parameters ModelConfig(model_name_or_path='deepseek-ai/DeepSeek-V2-Lite', model_revision='main', torch_dtype='bfloat16', trust_remote_code=True, attn_implementation='flash_attention_2', use_peft=False, lora_r=16, lora_alpha=32, lora_dropout=0.05, lora_target_modules=None, lora_modules_to_save=None, lora_task_type='CAUSAL_LM', use_rslora=False, load_in_8bit=False, load_in_4bit=False, bnb_4bit_quant_type='nf4', use_bnb_nested_quant=False)
+2026-03-21 20:15:46 - INFO - __main__ - Script parameters ScriptArguments(dataset_name='RoxanneWsyw/ESFT-summary', dataset_config=None, dataset_train_split='train', dataset_test_split='test', gradient_checkpointing_use_reentrant=False, ignore_bias_buffers=False)
+2026-03-21 20:15:46 - INFO - __main__ - Training parameters SFTConfig(
+_n_gpu=1,
+accelerator_config={'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False},
+adafactor=False,
+adam_beta1=0.9,
+adam_beta2=0.999,
+adam_epsilon=1e-08,
+attn_kl_weight=1.0,
+auto_find_batch_size=False,
+average_tokens_across_devices=False,
+batch_eval_metrics=False,
+benchmarks=[],
+bf16=True,
+bf16_full_eval=False,
+callbacks=[],
+chars_per_token=<CHARS_PER_TOKEN>,
+chat_template=None,
+cluster_mode=hierarchical-dynamic,
+cluster_num_groups=None,
+cluster_prune_ratio=None,
+cluster_prune_tau=1.0,
+data_seed=None,
+dataloader_drop_last=False,
+dataloader_num_workers=0,
+dataloader_persistent_workers=False,
+dataloader_pin_memory=True,
+dataloader_prefetch_factor=None,
+dataset_batch_size=None,
+dataset_kwargs=None,
+dataset_num_proc=None,
+dataset_text_field=text,
+ddp_backend=None,
+ddp_broadcast_buffers=None,
+ddp_bucket_cap_mb=None,
+ddp_find_unused_parameters=None,
+ddp_timeout=1800000000,
+debug=[],
+deepspeed=None,
+disable_teacher_dropout=True,
+disable_tqdm=False,
+dispatch_batches=None,
+do_eval=True,
+do_predict=False,
+do_train=False,
+entropy_slope_alpha=1.0,
+entropy_slope_beta=1.0,
+eval_accumulation_steps=None,
+eval_delay=0,
+eval_do_concat_batches=True,
+eval_on_start=False,
+eval_packing=None,
+eval_steps=None,
+eval_strategy=IntervalStrategy.NO,
+eval_use_gather_object=False,
+evaluation_strategy=None,
+fp16=False,
+fp16_backend=auto,
+fp16_full_eval=False,
+fp16_opt_level=O1,
+fsdp=[],
+fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False},
+fsdp_min_num_params=0,
+fsdp_transformer_layer_cls_to_wrap=None,
+full_determinism=False,
+gradient_accumulation_steps=1,
+gradient_checkpointing=True,
+gradient_checkpointing_kwargs={'use_reentrant': False},
+greater_is_better=None,
+group_by_length=False,
+half_precision_backend=auto,
+hub_always_push=False,
+hub_model_id=None,
+hub_model_revision=main,
+hub_private_repo=None,
+hub_strategy=HubStrategy.EVERY_SAVE,
+hub_token=<HUB_TOKEN>,
+ignore_data_skip=False,
+include_for_metrics=[],
+include_inputs_for_metrics=False,
+include_num_input_tokens_seen=False,
+include_tokens_per_second=False,
+jit_mode_eval=False,
+label_names=None,
+label_smoothing_factor=0.0,
+last_entropy_weight=1.0,
+layer_entropy_l1_layers=None,
+layer_entropy_l1_weight=1.0,
+learning_rate=1e-05,
+length_column_name=length,
+load_best_model_at_end=False,
+local_rank=0,
+log_level=info,
+log_level_replica=warning,
+log_on_each_node=True,
+logging_dir=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft/runs/Mar21_20-15-44_orchard-community-1,
+logging_first_step=False,
+logging_nan_inf_filter=True,
+logging_steps=1,
+logging_strategy=IntervalStrategy.STEPS,
+lr_scheduler_kwargs={'min_lr_rate': 0.1},
+lr_scheduler_type=SchedulerType.COSINE_WITH_MIN_LR,
+max_grad_norm=1.0,
+max_length=4096,
+max_seq_length=None,
+max_steps=-1,
+merging_metrics=None,
+metric_for_best_model=None,
+model_init_kwargs=None,
+mp_parameters=,
+neftune_noise_alpha=None,
+no_cuda=False,
+num_of_sequences=None,
+num_train_epochs=1,
+optim=OptimizerNames.ADAMW_TORCH,
+optim_args=None,
+optim_target_modules=None,
+output_dir=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft,
+overwrite_hub_revision=False,
+overwrite_output_dir=True,
+packing=False,
+past_index=-1,
+per_device_eval_batch_size=16,
+per_device_train_batch_size=2,
+prediction_loss_only=False,
+push_to_hub=False,
+push_to_hub_model_id=None,
+push_to_hub_organization=None,
+push_to_hub_revision=False,
+push_to_hub_token=<PUSH_TO_HUB_TOKEN>,
+ray_scope=last,
+remove_unused_columns=True,
+report_to=['wandb'],
+restore_callback_states_from_checkpoint=False,
+resume_from_checkpoint=None,
+router_manual_mask=None,
+router_prune_enable=True,
+router_prune_expert_per_layer=None,
+router_prune_interval=5,
+router_prune_min_keep=1,
+router_prune_start_step=None,
+router_prune_step_size=32,
+router_prune_use_plan=True,
+run_name=/project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft,
+save_on_each_node=False,
+save_only_model=False,
+save_safetensors=True,
+save_steps=500,
+save_strategy=SaveStrategy.NO,
+save_total_limit=None,
+seed=1234,
+skip_memory_metrics=True,
+split_batches=None,
+system_prompt=None,
+teacher_attn_implementation=None,
+teacher_model_name_or_path=None,
+teacher_model_revision=None,
+teacher_torch_dtype=auto,
+tf32=None,
+torch_compile=False,
+torch_compile_backend=None,
+torch_compile_mode=None,
+torch_empty_cache_steps=None,
+torchdynamo=None,
+tpu_metrics_debug=False,
+tpu_num_cores=None,
+use_cpu=False,
+use_ipex=False,
+use_legacy_prediction_loop=False,
+use_liger=False,
+use_liger_kernel=False,
+use_mps_device=False,
+wandb_entity=jayzxinkai-uc-san-diego,
+wandb_project=moe-honing,
+warmup_ratio=0.1,
+warmup_steps=0,
+weight_decay=0.0,
+weight_feature_rank=None,
+)
+2026-03-21 20:15:47 - INFO - datasets.builder - Found cached dataset esft-summary (/tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138)
+2026-03-21 20:15:47 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-d8c6f402c91a2432_*_of_00001.arrow
+2026-03-21 20:15:47 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-1975b68b541b14ab_*_of_00001.arrow
+2026-03-21 20:15:48 - INFO - __main__ - *** Initializing model kwargs ***
+2026-03-21 20:16:06 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-6435616c9a34cd0e_*_of_00001.arrow
+2026-03-21 20:16:06 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-264b8deb588933bd_*_of_00001.arrow
+2026-03-21 20:16:06 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-ad7de21f8027b38f_*_of_00001.arrow
+2026-03-21 20:16:06 - INFO - datasets.arrow_dataset - Loading cached processed dataset at /tmp/hf_cache/datasets/RoxanneWsyw___esft-summary/default/0.0.0/70d8f41993d8681cb5ccb26c656f4b9f5e0f8138/cache-545571e95ce7027c_*_of_00001.arrow
+2026-03-21 20:16:12 - INFO - __main__ - *** Train ***
+2026-03-21 20:16:12 - INFO - __main__ - DeepseekV2ForCausalLM(
+  (model): DeepseekV2Model(
+    (embed_tokens): Embedding(102400, 2048)
+    (layers): ModuleList(
+      (0): DeepseekV2DecoderLayer(
+        (self_attn): DeepseekV2FlashAttention2(
+          (q_proj): Linear(in_features=2048, out_features=3072, bias=False)
+          (kv_a_proj_with_mqa): Linear(in_features=2048, out_features=576, bias=False)
+          (kv_a_layernorm): DeepseekV2RMSNorm()
+          (kv_b_proj): Linear(in_features=512, out_features=4096, bias=False)
+          (o_proj): Linear(in_features=2048, out_features=2048, bias=False)
+          (rotary_emb): DeepseekV2YarnRotaryEmbedding()
+        )
+        (mlp): DeepseekV2MLP(
+          (gate_proj): Linear(in_features=2048, out_features=10944, bias=False)
+          (up_proj): Linear(in_features=2048, out_features=10944, bias=False)
+          (down_proj): Linear(in_features=10944, out_features=2048, bias=False)
+          (act_fn): SiLU()
+        )
+        (input_layernorm): DeepseekV2RMSNorm()
+        (post_attention_layernorm): DeepseekV2RMSNorm()
+      )
+      (1-26): 26 x DeepseekV2DecoderLayer(
+        (self_attn): DeepseekV2FlashAttention2(
+          (q_proj): Linear(in_features=2048, out_features=3072, bias=False)
+          (kv_a_proj_with_mqa): Linear(in_features=2048, out_features=576, bias=False)
+          (kv_a_layernorm): DeepseekV2RMSNorm()
+          (kv_b_proj): Linear(in_features=512, out_features=4096, bias=False)
+          (o_proj): Linear(in_features=2048, out_features=2048, bias=False)
+          (rotary_emb): DeepseekV2YarnRotaryEmbedding()
+        )
+        (mlp): DeepseekV2MoE(
+          (experts): ModuleList(
+            (0-63): 64 x DeepseekV2MLP(
+              (gate_proj): Linear(in_features=2048, out_features=1408, bias=False)
+              (up_proj): Linear(in_features=2048, out_features=1408, bias=False)
+              (down_proj): Linear(in_features=1408, out_features=2048, bias=False)
+              (act_fn): SiLU()
+            )
+          )
+          (gate): MoEGate()
+          (shared_experts): DeepseekV2MLP(
+            (gate_proj): Linear(in_features=2048, out_features=2816, bias=False)
+            (up_proj): Linear(in_features=2048, out_features=2816, bias=False)
+            (down_proj): Linear(in_features=2816, out_features=2048, bias=False)
+            (act_fn): SiLU()
+          )
+        )
+        (input_layernorm): DeepseekV2RMSNorm()
+        (post_attention_layernorm): DeepseekV2RMSNorm()
+      )
+    )
+    (norm): DeepseekV2RMSNorm()
+  )
+  (lm_head): Linear(in_features=2048, out_features=102400, bias=False)
+)
+2026-03-21 22:39:38 - INFO - __main__ - *** Save model ***
+2026-03-21 22:41:57 - INFO - __main__ - Model saved to /project/flame/haozeh/llm-honing/sft_models/deepseek-summary-sft

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddef2d3e003e6daaba60e27c15bbfbcec4a6fad8cab4a14f3376255328e32b50
+size 8504