Model save

README.md

@@ -1,5 +1,4 @@
 ---
-license: apache-2.0
 library_name: peft
 tags:
 - trl
@@ -7,7 +6,7 @@ tags:
 - generated_from_trainer
 datasets:
 - generator
-base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
+base_model: data/Tukan-1.1B-Chat-reasoning-sft_merged
 model-index:
 - name: Tukan-1.1B-Chat-reasoning-sft
   results: []
@@ -18,9 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Tukan-1.1B-Chat-reasoning-sft
 
-This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-Chat-v1.0](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0) on the generator dataset.
+This model was trained from scratch on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.2243
+- Loss: 1.1380
 
 ## Model description
 
@@ -54,10 +53,10 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.3384        | 0.24  | 10   | 1.2810          |
-| 1.2633        | 0.47  | 20   | 1.2418          |
-| 1.2495        | 0.71  | 30   | 1.2277          |
-| 1.2291        | 0.94  | 40   | 1.2244          |
+| 1.2298        | 0.24  | 10   | 1.1998          |
+| 1.1885        | 0.47  | 20   | 1.1653          |
+| 1.1674        | 0.71  | 30   | 1.1437          |
+| 1.1425        | 0.94  | 40   | 1.1381          |
 
 
 ### Framework versions

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af46676dbe82858b87ca1547c71c35fb5c75de6b10ae4a9ef49737c41e7ae06c
+oid sha256:6e7920098c524c2c57950f253babaf3c8f9f48ecf07eeb9c21fd35574bf96494
 size 26361536

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 0.99,
-    "eval_loss": 1.2242895364761353,
-    "eval_runtime": 24.7608,
+    "eval_loss": 1.1380302906036377,
+    "eval_runtime": 24.4133,
     "eval_samples": 662,
-    "eval_samples_per_second": 5.21,
-    "eval_steps_per_second": 1.737,
-    "train_loss": 1.282569306237357,
-    "train_runtime": 3226.4767,
+    "eval_samples_per_second": 5.284,
+    "eval_steps_per_second": 1.761,
+    "train_loss": 1.1835048993428547,
+    "train_runtime": 3217.2275,
     "train_samples": 25778,
-    "train_samples_per_second": 1.576,
+    "train_samples_per_second": 1.58,
     "train_steps_per_second": 0.013
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 0.99,
-    "eval_loss": 1.2242895364761353,
-    "eval_runtime": 24.7608,
+    "eval_loss": 1.1380302906036377,
+    "eval_runtime": 24.4133,
     "eval_samples": 662,
-    "eval_samples_per_second": 5.21,
-    "eval_steps_per_second": 1.737
+    "eval_samples_per_second": 5.284,
+    "eval_steps_per_second": 1.761
 }

runs/Jan20_09-47-06_98f107f1aa39/events.out.tfevents.1705747412.98f107f1aa39.59272.2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b251b451f30eebb65f8c6f2e9add2583d9865c3ab0b2e7e8860eda8ce441ef33
-size 6977
+oid sha256:be2a18398cffadd3c9d36fe951c2e8dc0816f24d9afcb29c43d3579da8210556
+size 7325

runs/Jan20_09-47-06_98f107f1aa39/events.out.tfevents.1705750653.98f107f1aa39.59272.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49057e8ddadd2ef376f5408586cca18ce2b101f1754327f775ac419cfd347940
+size 354

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 0.99,
-    "train_loss": 1.282569306237357,
-    "train_runtime": 3226.4767,
+    "train_loss": 1.1835048993428547,
+    "train_runtime": 3217.2275,
     "train_samples": 25778,
-    "train_samples_per_second": 1.576,
+    "train_samples_per_second": 1.58,
     "train_steps_per_second": 0.013
 }

trainer_state.json

@@ -11,96 +11,96 @@
     {
       "epoch": 0.02,
       "learning_rate": 4.9930094929529506e-05,
-      "loss": 1.458,
+      "loss": 1.2348,
       "step": 1
     },
     {
       "epoch": 0.12,
       "learning_rate": 4.827184371610511e-05,
-      "loss": 1.4238,
+      "loss": 1.2258,
       "step": 5
     },
     {
       "epoch": 0.24,
       "learning_rate": 4.332629679574566e-05,
-      "loss": 1.3384,
+      "loss": 1.2298,
       "step": 10
     },
     {
       "epoch": 0.24,
-      "eval_loss": 1.2809978723526,
-      "eval_runtime": 24.7617,
-      "eval_samples_per_second": 5.21,
-      "eval_steps_per_second": 1.737,
+      "eval_loss": 1.1997809410095215,
+      "eval_runtime": 24.4233,
+      "eval_samples_per_second": 5.282,
+      "eval_steps_per_second": 1.761,
       "step": 10
     },
     {
       "epoch": 0.35,
       "learning_rate": 3.5847093477938956e-05,
-      "loss": 1.2865,
+      "loss": 1.2091,
       "step": 15
     },
     {
       "epoch": 0.47,
       "learning_rate": 2.686825233966061e-05,
-      "loss": 1.2633,
+      "loss": 1.1885,
       "step": 20
     },
     {
       "epoch": 0.47,
-      "eval_loss": 1.2418025732040405,
-      "eval_runtime": 24.7653,
-      "eval_samples_per_second": 5.209,
-      "eval_steps_per_second": 1.736,
+      "eval_loss": 1.1652644872665405,
+      "eval_runtime": 24.4242,
+      "eval_samples_per_second": 5.282,
+      "eval_steps_per_second": 1.761,
       "step": 20
     },
     {
       "epoch": 0.59,
       "learning_rate": 1.7631120639727393e-05,
-      "loss": 1.2433,
+      "loss": 1.1647,
       "step": 25
     },
     {
       "epoch": 0.71,
       "learning_rate": 9.412754953531663e-06,
-      "loss": 1.2495,
+      "loss": 1.1674,
       "step": 30
     },
     {
       "epoch": 0.71,
-      "eval_loss": 1.2276524305343628,
-      "eval_runtime": 24.7715,
-      "eval_samples_per_second": 5.208,
-      "eval_steps_per_second": 1.736,
+      "eval_loss": 1.1437482833862305,
+      "eval_runtime": 24.4259,
+      "eval_samples_per_second": 5.281,
+      "eval_steps_per_second": 1.76,
       "step": 30
     },
     {
       "epoch": 0.83,
       "learning_rate": 3.3493649053890326e-06,
-      "loss": 1.2364,
+      "loss": 1.151,
       "step": 35
     },
     {
       "epoch": 0.94,
       "learning_rate": 2.7922934437178695e-07,
-      "loss": 1.2291,
+      "loss": 1.1425,
       "step": 40
     },
     {
       "epoch": 0.94,
-      "eval_loss": 1.2243515253067017,
-      "eval_runtime": 24.7908,
-      "eval_samples_per_second": 5.204,
-      "eval_steps_per_second": 1.735,
+      "eval_loss": 1.1380583047866821,
+      "eval_runtime": 24.4167,
+      "eval_samples_per_second": 5.283,
+      "eval_steps_per_second": 1.761,
       "step": 40
     },
     {
       "epoch": 0.99,
       "step": 42,
       "total_flos": 6.447644673468006e+16,
-      "train_loss": 1.282569306237357,
-      "train_runtime": 3226.4767,
-      "train_samples_per_second": 1.576,
+      "train_loss": 1.1835048993428547,
+      "train_runtime": 3217.2275,
+      "train_samples_per_second": 1.58,
       "train_steps_per_second": 0.013
     }
   ],