Add policy init lora checkpoint

README.md

@@ -1,3 +1,9 @@
----
-license: apache-2.0
----
+---
+library_name: peft
+---
+## Training procedure
+
+### Framework versions
+
+
+- PEFT 0.4.0

all_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 1.417236328125,
+    "eval_runtime": 341.5207,
+    "eval_samples_per_second": 2.998,
+    "eval_steps_per_second": 0.047,
+    "train_loss": 1.4022162783462389,
+    "train_runtime": 13253.0183,
+    "train_samples_per_second": 2.186,
+    "train_steps_per_second": 0.017
+}

config.json

@@ -0,0 +1,45 @@
+{
+  "_name_or_path": "/dataset/llms/LLaVA_RLHF/LLaVA_Video-RLHF/pretrained/llava-v1.5-13b-lora_r32a32_sftv2_short2long_lr1e-4_merged",
+  "architectures": [
+    "LlavaLlamaForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "cache_shape": null,
+  "eos_token_id": 2,
+  "freeze_mm_mlp_adapter": false,
+  "freeze_mm_vision_resampler": false,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "image_aspect_ratio": "pad",
+  "image_grid_pinpoints": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 13824,
+  "max_length": 4096,
+  "max_position_embeddings": 4096,
+  "mm_hidden_size": 1024,
+  "mm_projector_lr": 2e-05,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+  "model_type": "llava",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 40,
+  "pad_token_id": 0,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.31.0",
+  "tune_mm_mlp_adapter": false,
+  "tune_mm_vision_resampler": false,
+  "unfreeze_mm_vision_tower": false,
+  "use_cache": false,
+  "use_mm_proj": true,
+  "vocab_size": 32000
+}

eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 1.417236328125,
+    "eval_runtime": 341.5207,
+    "eval_samples_per_second": 2.998,
+    "eval_steps_per_second": 0.047
+}

lora_default/adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "/dataset/llms/LLaVA_RLHF/LLaVA_Video-RLHF/pretrained/llava-v1.5-13b-lora_r32a32_sftv2_short2long_lr1e-4_merged",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 64,
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "v_proj",
+    "o_proj",
+    "gate_proj",
+    "up_proj",
+    "down_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

lora_default/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38464449a6aa9803c0975e6108c31918a3cea99c784856ba0fcb796ced6ed1da
+size 500889293

metrics.json

@@ -0,0 +1 @@
+{"run_name": "/dataset/llms/LLaVA_RLHF/LLaVA_Video-RLHF/pretrained/LLaVA_Video-RL-INIT-13bv2s2l_dv1-v1.5-336-lora-padding", "train_runtime": 13253.0183, "train_samples_per_second": 2.186, "train_steps_per_second": 0.017, "train_loss": 1.4022162783462389, "epoch": 1.0, "eval_loss": 1.417236328125, "eval_runtime": 341.5207, "eval_samples_per_second": 2.998, "eval_steps_per_second": 0.047}

non_lora_trainables.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc6b9cc23771073378c3887f3524986a3b61cac85cbc085f322b318241bc6845
+size 455

runs/Feb08_05-47-04_gpu-1/events.out.tfevents.1707372728.gpu-1.16464.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31f495c1ba8a65ec486f8fdd5755f6200c89cdff133aa5b69826beb75604fdf8
+size 15417

runs/Feb08_05-47-04_gpu-1/events.out.tfevents.1707386324.gpu-1.16464.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76d16c829a3772973def0f1ede39c028b6f9a663191c180688feee58f8fbd35a
+size 359

train_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 1.0,
+    "train_loss": 1.4022162783462389,
+    "train_runtime": 13253.0183,
+    "train_samples_per_second": 2.186,
+    "train_steps_per_second": 0.017
+}