update model checkpoints

README.md

@@ -0,0 +1,76 @@
+---
+library_name: peft
+tags:
+- alignment-handbook
+- generated_from_trainer
+datasets:
+- llama-duo/synth_summarize_dataset_dedup
+base_model: google/gemma-7b
+model-index:
+- name: gemma7b-summarize-gpt4o-256k
+  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. -->
+
+# gemma7b-summarize-gpt4o-256k
+
+This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the llama-duo/synth_summarize_dataset_dedup dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.4681
+
+## 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: 0.0002
+- train_batch_size: 4
+- eval_batch_size: 2
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 64
+- total_eval_batch_size: 16
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 10
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 1.0805        | 1.0   | 439  | 2.5210          |
+| 0.9397        | 2.0   | 878  | 2.4361          |
+| 0.8628        | 3.0   | 1317 | 2.4056          |
+| 0.8131        | 4.0   | 1756 | 2.4177          |
+| 0.7788        | 5.0   | 2195 | 2.4166          |
+| 0.771         | 6.0   | 2634 | 2.4329          |
+| 0.7459        | 7.0   | 3073 | 2.4458          |
+| 0.745         | 8.0   | 3512 | 2.4609          |
+| 0.7433        | 9.0   | 3951 | 2.4640          |
+| 0.7376        | 10.0  | 4390 | 2.4681          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.40.0
+- Pytorch 2.1.2+cu121
+- Datasets 2.18.0
+- Tokenizers 0.19.1

adapter_config.json

@@ -10,23 +10,18 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 16,
+  "lora_alpha": 64,
   "lora_dropout": 0.05,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 8,
+  "r": 32,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "up_proj",
-    "v_proj",
-    "gate_proj",
-    "down_proj",
-    "o_proj",
     "q_proj",
-    "k_proj"
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:74b9a6f63b5c4eedb5848c31c32ac5684bc2b59e6653085164bfb60954b9e23c
-size 50056096
+oid sha256:8759c0a85b4745cee2f460373cf01029db543f5b9b6a3c130589ab135b10938e
+size 25705248

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 10.0,
+    "eval_loss": 2.4680683612823486,
+    "eval_runtime": 0.2341,
+    "eval_samples": 25,
+    "eval_samples_per_second": 42.712,
+    "eval_steps_per_second": 4.271,
+    "total_flos": 1.3402520949471838e+19,
+    "train_loss": 1.3662878394941533,
+    "train_runtime": 10620.987,
+    "train_samples": 258442,
+    "train_samples_per_second": 26.444,
+    "train_steps_per_second": 0.413
+}

config.json

@@ -0,0 +1,44 @@
+{
+  "_name_or_path": "google/gemma-7b",
+  "architectures": [
+    "GemmaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 2,
+  "eos_token_id": 1,
+  "head_dim": 256,
+  "hidden_act": "gelu",
+  "hidden_activation": null,
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 24576,
+  "max_position_embeddings": 8192,
+  "model_type": "gemma",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 16,
+  "pad_token_id": 0,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": false,
+    "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,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.40.0",
+  "use_cache": true,
+  "vocab_size": 256000
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 10.0,
+    "eval_loss": 2.4680683612823486,
+    "eval_runtime": 0.2341,
+    "eval_samples": 25,
+    "eval_samples_per_second": 42.712,
+    "eval_steps_per_second": 4.271
+}

runs/Jun13_11-37-34_gpu1-2/events.out.tfevents.1718250264.gpu1-2.1350172.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:122f64777a6ab4864b4cddb8d796c9576a99869afad2c7f2571736fdec336106
+size 193884

runs/Jun13_11-37-34_gpu1-2/events.out.tfevents.1718260886.gpu1-2.1350172.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65219c7f29b3ea660fde7d4a10fe0d43ab74f7b4a55884badb3d97db9da05b7d
+size 359

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 1.3402520949471838e+19,
+    "train_loss": 1.3662878394941533,
+    "train_runtime": 10620.987,
+    "train_samples": 258442,
+    "train_samples_per_second": 26.444,
+    "train_steps_per_second": 0.413
+}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:24a383cf6f925740a20f6dd6ba70e8d3894ed9b88a43cd4b885b70d11c4195c8
+oid sha256:90337f4fd44f4a744a1a87fd6709de0ba38f95b6287b7d28c0488029993945e5
 size 5176