Training in progress, epoch 2, checkpoint

checkpoint-500/adapter_config.json

@@ -1,6 +1,6 @@
 {
   "auto_mapping": null,
-  "base_model_name_or_path": "Qwen/Qwen1.5-4B",
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,

checkpoint-500/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:943b1d7051b23585008987d2755a120ce1aafe2a15c9ab026bf79e1a1207f6e7
+size 143269386

checkpoint-500/added_tokens.json

@@ -1,5 +1,5 @@
 {
-  "<|endoftext|>": 151643,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644
+  "</s>": 2,
+  "<s>": 1,
+  "<unk>": 0
 }

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2451aef11e0231977b0c6bd4b3acb406bd19456b151c65e3726043fbae73da4d
-size 224537202
+oid sha256:54af755e84ffff4b5f22c3367b3e6c0077ecef064d843bcfb661b4cb464c3439
+size 286585234

checkpoint-500/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee4a99ed0a61a725eea7c7573b5ee2b0f6cda717255e14762f24f5d5f8004a2e
+oid sha256:637071e2c166d5fdc9e66939a1ce0281249926678e09dda5c2a3837705f547b4
 size 15024

checkpoint-500/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5bf67fe471475a8760c25bd15679bc7a71860047dfc2890444e3745d53c56bc0
+oid sha256:ab331e19da81b1a266c0446866904296ba1cd7b8d8351d419f665735a1513b85
 size 15024

checkpoint-500/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6a5e94f8ec9a39cd10456daafe0d398786ce16bb44c57339919b3ae8f515a3c
+oid sha256:1c175aee262e2f0573545a1cb2c24c1865d88acea5f14929978a61d055492e51
 size 15024

checkpoint-500/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4fb6e01e28099ff0bb08b377ea619c51ab525facc064f6be44e2657402c60e5c
+oid sha256:2c3e72cdbefc33d1d8ac8866ffacf28f990340106779067c64a3117cd8e4a1e1
 size 15024

checkpoint-500/special_tokens_map.json

@@ -1,14 +1,6 @@
 {
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "eos_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": "<|endoftext|>"
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "</s>",
+  "unk_token": "<unk>"
 }

checkpoint-500/tokenizer.model

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

checkpoint-500/tokenizer_config.json

@@ -1,24 +1,23 @@
 {
-  "add_prefix_space": false,
   "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
+    "0": {
+      "content": "<unk>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151644": {
-      "content": "<|im_start|>",
+    "1": {
+      "content": "<s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151645": {
-      "content": "<|im_end|>",
+    "2": {
+      "content": "</s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -26,18 +25,16 @@
       "special": true
     }
   },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
-  "eos_token": "<|endoftext|>",
-  "errors": "replace",
-  "model_max_length": 32768,
-  "pad_token": "<|endoftext|>",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
+  "eos_token": "</s>",
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "left",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true
 }

checkpoint-500/trainer_state.json

@@ -10,89 +10,70 @@
   "log_history": [
     {
       "epoch": 0.4,
-      "grad_norm": 0.6251392960548401,
       "learning_rate": 5e-05,
-      "loss": 2.427,
+      "loss": 2.1914,
       "step": 100
     },
     {
       "epoch": 0.8,
-      "grad_norm": 0.6590011715888977,
       "learning_rate": 5e-05,
-      "loss": 2.2748,
+      "loss": 1.798,
       "step": 200
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5142539682539683,
-      "eval_loss": 2.336690902709961,
-      "eval_runtime": 5.4733,
-      "eval_samples_per_second": 91.353,
-      "eval_steps_per_second": 11.51,
+      "eval_accuracy": 0.6067341772151899,
+      "eval_loss": 1.8213441371917725,
+      "eval_runtime": 4.8116,
+      "eval_samples_per_second": 103.916,
+      "eval_steps_per_second": 13.093,
       "step": 250
     },
     {
       "epoch": 1.0,
-      "eval_exact_match": 16.8,
-      "eval_f1": 24.850246583481884,
+      "eval_exact_match": 25.4,
+      "eval_f1": 35.50992875545509,
       "step": 250
     },
     {
       "epoch": 1.2,
-      "grad_norm": 0.8194239139556885,
       "learning_rate": 5e-05,
-      "loss": 2.2322,
+      "loss": 1.7534,
       "step": 300
     },
     {
       "epoch": 1.6,
-      "grad_norm": 0.9867294430732727,
       "learning_rate": 5e-05,
-      "loss": 2.1904,
+      "loss": 1.7163,
       "step": 400
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.9660400152206421,
       "learning_rate": 5e-05,
-      "loss": 2.1657,
+      "loss": 1.7,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5168253968253969,
-      "eval_loss": 2.323688507080078,
-      "eval_runtime": 5.1683,
-      "eval_samples_per_second": 96.744,
-      "eval_steps_per_second": 12.19,
+      "eval_accuracy": 0.6077215189873417,
+      "eval_loss": 1.8046445846557617,
+      "eval_runtime": 5.1228,
+      "eval_samples_per_second": 97.603,
+      "eval_steps_per_second": 12.298,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 17.2,
-      "eval_f1": 24.620036075036083,
+      "eval_exact_match": 25.8,
+      "eval_f1": 35.76454372424963,
       "step": 500
     }
   ],
   "logging_steps": 100,
   "max_steps": 12500,
-  "num_input_tokens_seen": 0,
   "num_train_epochs": 50,
   "save_steps": 500,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": false
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 1.2935386007339008e+16,
-  "train_batch_size": 1,
+  "total_flos": 3.1895934673616896e+16,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ede3565a5658e9b8743d6efdd527fd23709623a130469c81f1191d07c33fbcb1
-size 5304
+oid sha256:ddd08619aa8955df0f518864a252643438869f6968a9f390d3026fbeb48e7ddf
+size 4728